65+ PagerDuty Interview Questions: Integrations (2027)
76 PagerDuty integration interview questions: Alertmanager, Datadog, CloudWatch, Azure, Slack, Jira, ServiceNow, change events, webhooks, Terraform.
PagerDuty integration interviews test whether you can connect real monitoring and delivery tools to PagerDuty and keep the pages useful: Alertmanager, Grafana, Datadog, CloudWatch, Azure Monitor, Slack, Jira, ServiceNow, change events from CI/CD, webhooks and Terraform. This page has 76 questions with direct answers, grouped by topic, followed by troubleshooting scenarios.
Facts about integration steps, endpoints, roles and limits were checked against PagerDuty, Prometheus, Microsoft and AWS documentation in November 2026. Where a product page was not reachable or its wording was unclear, the answer says so instead of guessing.
Key takeaways
- Every integration answers three questions: how does an event get in (a key or URL), how does it get routed (service or orchestration) and how does it close (a resolve event or a sync).
- Alertmanager uses
routing_keyfor Events API v2 integrations andservice_keyfor the older Prometheus integration type. Resolve messages are sent by default (send_resolvedis true). - A CloudWatch alarm reaches PagerDuty through an SNS topic with an HTTPS subscription, with raw message delivery left off, and an OK notification on the same topic resolves the incident.
- Change events from your pipeline give responders context but never page anyone.
- Webhooks v3 are the current webhook version. Version 2 has been in end-of-support since 31 October 2022.
- Manage services, integrations, Slack connections, webhook subscriptions and orchestration rules with the PagerDuty Terraform provider so changes are reviewed.
How this page fits with the other PagerDuty guides
This page is about connecting PagerDuty to other tools. For the core object model, Events API v2 fields, deduplication and escalation limits, read the main set, PagerDuty engineer interview questions and answers. For incident drills use scenario-based PagerDuty questions and for a study plan see the PagerDuty interview preparation guide. To compare PagerDuty with other products, read 12 on-call management tools for SRE and DevOps.
| Source | How it reaches PagerDuty | How it resolves |
|---|---|---|
| Prometheus Alertmanager | Receiver with routing_key (Events API v2) or service_key | Resolve notification when the alert clears |
| Grafana alerting | PagerDuty contact point with an Events API v2 integration key | Resolved alert state |
| Datadog | PagerDuty integration tile, then @pagerduty-ServiceName in a monitor | Datadog sends a resolve when the metric recovers |
| Amazon CloudWatch | Alarm to SNS topic to HTTPS subscription on the PagerDuty URL | OK notification on the same topic |
| Azure Monitor | Action group with a webhook action to the PagerDuty URL | Depends on the alert payload and integration |
| CI/CD pipeline | Change events endpoint | Not applicable (no incident) |
| Slack, Jira, ServiceNow | Native extensions configured in PagerDuty | Two-way status sync |
Integration basics
1. What is a PagerDuty integration?
An integration is the connection that lets an external tool send events to a PagerDuty service or orchestration, or lets PagerDuty exchange data with another system. Inbound integrations use an integration key or URL. Extensions such as Slack, Jira and ServiceNow work in both directions.
2. What is the difference between a service integration key and an Event Orchestration key?
A service integration key sends events straight to one service. An orchestration key sends them into an Event Orchestration, which can route them to different services by rules. For orchestration, PagerDuty documents the URL form https://events.pagerduty.com/x-ere/[YOUR_INTEGRATION_KEY_HERE] for tools that need a URL, such as CloudWatch.
3. Which integration type do you pick for a tool PagerDuty does not list?
Use a generic Events API v2 integration and have the tool call it with a webhook or script. Events API v2 is the general way to send alert events, and the same integration key also accepts change events. Check whether the tool lets you template the JSON body.
4. How do Alertmanager and PagerDuty connect?
Add a pagerduty_configs block to a receiver with the key, route the right alerts to that receiver, and restart or reload Alertmanager. PagerDuty's guide offers two paths: send to an Event Orchestration or add an integration directly on a service.
receivers:
- name: pd-critical
pagerduty_configs:
- routing_key_file: /etc/alertmanager/pd-critical.key
severity: critical
- name: pd-warning
pagerduty_configs:
- routing_key_file: /etc/alertmanager/pd-warning.key
severity: warning
route:
receiver: pd-warning
group_by: [alertname, cluster, service]
routes:
- matchers:
- severity = critical
receiver: pd-critical
5. What is the difference between routing_key and service_key in Alertmanager?
routing_key is for an Events API v2 integration. service_key is for the Prometheus integration type, which uses the older Events API v1. PagerDuty's guide says that if you set routing_key with v2, the integration behind the key must be of type Events API v2. The two fields are mutually exclusive with their _file variants.
6. How do you keep the Alertmanager key out of the config file?
Use routing_key_file (or service_key_file) to read the key from a file mounted from your secret store. In Kubernetes that is usually a Secret mounted into the Alertmanager Pod. Never commit keys to Git.
7. How does Grafana alerting page PagerDuty?
Create a service with an Events API V2 integration in PagerDuty, copy the integration key, then in Grafana add a contact point of type PagerDuty with that key and select it in the alert rule's notification settings. Grafana's Test button applies to the Grafana Alertmanager only.
8. How does Datadog trigger a PagerDuty incident?
Install the PagerDuty integration in Datadog, authorize it, and register each PagerDuty service with its name and integration key. A monitor then notifies @pagerduty-ServiceName in its message or notification list. Posting @pagerduty-ServiceName in the event stream also creates an incident, which is a quick test.
9. How do Datadog monitors resolve in PagerDuty?
When the metric returns to its range, Datadog sends a resolve event and the PagerDuty incident closes. You can also comment @pagerduty-resolve or @pagerduty-acknowledge on the original event stream post. Those comments appear in the incident's Timeline.
10. How does a CloudWatch alarm reach PagerDuty?
Create a Standard SNS topic, add an HTTPS subscription with the PagerDuty integration URL as the endpoint, confirm it, and set the alarm's In alarm action to that topic. PagerDuty can receive it either directly on a service or through an Event Orchestration URL. The integration expects a nested JSON-encoded object in the Message property.
11. Why must raw message delivery be off on the SNS subscription?
PagerDuty's CloudWatch guide tells you to leave Enable raw message delivery unchecked, because the integration reads the standard SNS envelope with the JSON message nested inside. With raw delivery on, the payload is not in the expected shape and no alert triggers.
12. How do CloudWatch incidents resolve automatically?
Add an OK notification on the alarm that publishes to the same SNS topic. When the metric returns to normal, CloudWatch publishes the OK message, and the integration resolves the incident. Without it, someone resolves the incident by hand. For alarm design, see how to automate cloud alerts with Amazon CloudWatch.
Cloud, chat and ticketing integrations
13. How does Azure Monitor send alerts to PagerDuty?
Create a service with the Microsoft Azure integration type in PagerDuty and copy its integration URL. In Azure Monitor, edit or create an action group and add a Webhook action with that URL. Microsoft's own walkthrough covers Service Health alerts, so test with the alert type you actually use.
14. What do you verify after wiring Azure Monitor?
Fire a test notification from the action group, then confirm an incident appears on the right service with a clear title, and that the recovery behaves as you expect. Check that the action group is attached to the alert rules you care about. An unattached action group sends nothing.
15. What does PagerDuty's Amazon EventBridge integration do?
It exposes PagerDuty as an EventBridge partner event source. You name the event source on the service integration, then associate it with an event bus in AWS. The partner source name follows aws.partner/pagerduty.com/{pagerduty-subdomain}-{eventSourceName}. Event patterns in EventBridge then filter the PagerDuty events you want.
16. How would you start an automated AWS fix from a PagerDuty incident?
Send PagerDuty events to an EventBridge bus, write a rule that matches the incident event, and target an automation. AWS documents that Systems Manager Automation is supported as an EventBridge rule target, with approval steps and rate controls available in the runbook. Keep human approval on anything destructive.
17. Does PagerDuty integrate with AWS events beyond CloudWatch?
PagerDuty publishes integration guides for AWS sources such as the AWS Health Dashboard and AWS Security Hub in addition to CloudWatch and EventBridge. Check the integration directory for the current list before you design around one.
18. What can responders do from Slack?
Declare, acknowledge, escalate and resolve incidents, run Incident Workflows, send status updates and assign roles, all without leaving Slack. Actions sync both ways, so a change in Slack updates the incident in PagerDuty and the reverse. Each user must link their PagerDuty and Slack accounts, and actions follow their PagerDuty permissions.
19. What are the Slack channel types?
PagerDuty describes dedicated incident channels (created per incident with details pinned), notification channels (updates by rule, for example by service, team or priority) and announcement channels (broadcast to stakeholders). Each can be set to highlights only, status updates only, all updates or no updates.
20. How do you connect a service to a Slack channel?
Use /pd connect in Slack or the Slack configuration page under Configuration, Integrations, Extensions. Pick the PagerDuty account, the service or team, and a notification type: Responder (with action buttons) or Stakeholder (without). After that, an incident card posts to the channel whenever an incident is declared.
21. Who can set up the Slack integration?
Initial workspace mapping needs an Admin, Global Admin or Account Owner base role in PagerDuty, and installing the Slack app needs a Slack workspace admin or owner. Team Managers, Admins and Account Owners can manage channel connections, while Responders have view-only access.
22. Why can a Slack user not declare an incident?
Usually their Slack account is not linked to a PagerDuty user. PagerDuty states that unlinked users cannot declare incidents unless the setting that allows any Slack user to declare incidents on services is enabled. Link accounts or decide deliberately whether to enable that setting.
23. What does the Jira Cloud integration do?
It creates a Jira issue from a PagerDuty incident, either with a Create Jira Issue action or automatically, and keeps status in sync both ways using a mapping you configure. It can also sync priority and notes. A PagerDuty user license is required.
24. How do you create PagerDuty incidents from Jira?
Create them manually from an issue, or use a JQL rule: tick the option to create an incident if the issue matches your JQL statement and enter the query. Use a narrow JQL (specific project, type and priority) so ordinary tickets do not page anyone.
25. What limits does the Jira Cloud integration have?
PagerDuty's FAQ says only one Jira issue is created per incident, and the workflow action allows one Jira Cloud instance mapped to one workflow mapping. Jira Server and Data Center use a different integration guide.
26. What does the ServiceNow integration sync?
The current guide covers the ServiceNow v8 integration. It creates a linked ServiceNow incident from a PagerDuty incident, automatically or manually, and keeps trigger, reassignment and resolution in sync. Priority syncs once mapping is configured, and ServiceNow change requests can be sent to PagerDuty as change events. It requires a Business or Enterprise PagerDuty plan.
Change events and CI/CD
27. What is the change events endpoint and what is required?
Post JSON to https://events.pagerduty.com/v2/change/enqueue. The required fields are routing_key, a 32-character integration key (the same key works for alert and change events), and payload.summary, up to 1024 characters. Source, timestamp, custom details and links are optional but make the event useful.
28. Can you send change events to an Event Orchestration key?
No. PagerDuty's overview says change events cannot be sent to ruleset integration keys, which begin with R. Use a service integration key for change events and keep orchestration keys for alerts.
29. How do you send a change event from a CI job?
Call the endpoint with curl after the deploy step succeeds, using a secret for the key.
curl -s -X POST https://events.pagerduty.com/v2/change/enqueue \
-H "Content-Type: application/json" \
-d '{
"routing_key": "'"$PD_SERVICE_KEY"'",
"payload": {
"summary": "Deployed checkout-api 3f9c2ab to production",
"source": "ci.example.com",
"timestamp": "2026-11-12T09:30:00Z",
"custom_details": { "build_number": "1482", "run_time": "6m12s" }
},
"links": [{ "href": "https://ci.example.com/job/1482", "text": "Build log" }]
}'
30. How do you send change events from GitHub Actions?
PagerDuty publishes a GitHub Action for change events that needs a service with a GitHub integration and the integration key in a repository secret. It handles push and pull request events, and for pull requests it creates an event only when the PR is merged. Pick one trigger type, because using both can create two events for one change. Check the repository's activity before depending on it.
31. Is there an Ansible option?
Yes. The community.general.pagerduty_change module tracks a code or infrastructure change as a change event. A new event is created each time the task runs, so add it as one task at the end of a play, not inside a loop.
32. Why send change events at all?
PagerDuty's overview says a change event does not notify anyone. It adds context for triage on service activity streams and incident details pages. "What changed in the last hour" is the first question in most incidents, and this answers it without hunting through pipelines.
33. What makes a change event useful?
A summary that names the service and version, a source that identifies the pipeline, an accurate timestamp, custom details such as build number and commit, and a link to the build or pull request. Send one event per production change, not one per pipeline stage.
34. How would you tie Kubernetes rollouts to change events?
Run kubectl rollout status in the pipeline, and send the change event only after it succeeds, so the event reflects a completed change. On failure, trigger your rollback and send a second change event describing it. See the role of Kubernetes in a DevOps pipeline for the pipeline steps.
Kubernetes and Prometheus patterns
35. Which Kubernetes conditions deserve a page?
Page on symptoms that hurt users: a workload with no ready replicas, sustained error rate, a node pool out of capacity, a failing control plane. Send crash loops and restarts to a chat or low-urgency route unless they affect availability. Lists of useful alerts are in 10 monitoring alerts every DevOps team should configure.
36. Write a crash loop alert rule.
kube-state-metrics exposes kube_pod_container_status_waiting_reason with a reason label, so a rule can match CrashLoopBackOff.
groups:
- name: pods
rules:
- alert: PodCrashLooping
expr: max_over_time(kube_pod_container_status_waiting_reason{reason="CrashLoopBackOff"}[5m]) >= 1
for: 15m
labels:
severity: warning
annotations:
summary: "{{ $labels.namespace }}/{{ $labels.pod }} is crash looping"
37. How does an Alertmanager alert get its PagerDuty severity?
The receiver's severity field sets it, and the default is error. Events API v2 accepts critical, error, warning and info. The simplest reliable design is one receiver per severity, as in question 4, and routes that match on your alert label. PagerDuty's guide says the integration can create high or low urgency incidents based on that severity.
38. Why does group_by matter for PagerDuty?
Alertmanager groups alerts that share the listed label values into one notification. Without sensible grouping, one outage in a cluster pages once per Pod. Group by alert name plus the dimension an engineer fixes, usually cluster and service. A group_by of ['...'] disables aggregation, which is rarely what you want for paging.
39. What are the Alertmanager timing defaults?
group_wait is 30 seconds, group_interval is 5 minutes and repeat_interval is 4 hours. Child routes inherit them from the parent. A short group_wait pages sooner but sends more partial groups, and a long repeat_interval avoids re-notifying about an alert someone already acknowledged.
40. How does inhibition reduce pages?
An inhibit rule mutes target alerts while source alerts are firing, for labels that match in both. For example, mute warning alerts for a cluster while a ClusterDown alert is active. Use the source_matchers, target_matchers and equal fields. The older *_match fields are deprecated.
41. Why might a PagerDuty incident resolve late after Alertmanager clears an alert?
PagerDuty's guide says incidents resolve automatically unless send_resolved is false, and that resolve notifications may wait until the next group_interval. Delivery is best effort. So a five-minute lag is normal with the default interval, and a missing resolve usually means send_resolved: false.
42. How do you run Alertmanager in high availability with PagerDuty?
Run several Alertmanager instances as a cluster with the --cluster-* flags so silences and notification state replicate. The Alertmanager documentation says Prometheus should list all Alertmanagers as targets rather than load balance across them. The cluster handles deduplication, so PagerDuty should not see duplicate pages.
43. How do you prove the whole alert path works?
Run an always-firing alert end to end, often named Watchdog (Alertmanager's own examples use alertname = Watchdog). Route it to a heartbeat check or a low-urgency service and alarm if it stops arriving. That detects a broken Prometheus, Alertmanager or integration key before a real incident does.
44. How do you test a new integration safely?
Point it at a test service with a notification rule that only reaches you, send a test event, verify routing, urgency, resolve and links, then switch the integration to the production service. Keep a maintenance window or a low-urgency escalation policy in place while testing.
Event Orchestration as routing glue
45. How do global and service orchestration differ?
Global Orchestrations run before routing and apply to matching events whichever service they reach. Service Orchestrations run after a routing rule, on one service. When rules conflict, the rule listed highest wins in a set, and service-level actions can override global ones.
46. What can an orchestration rule do?
Set severity (info, error, warning, critical), set priority (if priority is enabled on the account), suppress an alert, annotate the incident, pause notifications for a number of seconds, override the escalation policy (advanced tier), fire webhooks, and run Automation Actions on service orchestrations. A ruleset holds up to 25 rules.
47. What does pause notifications do?
It holds the alert in a suspended state for the number of seconds you set before it triggers, and the alert can also be triggered or resolved manually in the meantime. It suits flapping signals, at the cost of delaying a real problem by the same window.
48. How do you make urgency follow event severity?
Set severity in the integration or orchestration, then configure the service so urgency is derived from severity. PagerDuty states orchestration severity can drive dynamic notifications and urgency. Map critical to high urgency and warning to low, and test with sample events.
49. What are cache variables used for?
Cache variables remember values from earlier events so later rules can use them, and can be referenced with double curly braces in custom fields and webhook JSON bodies. They must exist before you reference them. Use them for patterns like counting repeats, not for permanent storage.
50. Which orchestration features depend on the plan?
PagerDuty's documentation ties Global Orchestrations to the AIOps add-on (or Signal Intelligence) and the Advanced tier, and Automation Actions to the Automation Actions add-on. Service Orchestrations are available on all packages. Confirm the entitlements before promising a design in an interview or to a team.
Webhooks v3
51. What are webhooks v3 and how do you create one?
Webhooks v3 is PagerDuty's current webhook version. You create a webhook subscription through the Webhook Subscriptions API. Each subscription has event types, a filter and an HTTP delivery method with a destination URL.
52. What filters can a subscription use?
A service reference, a team reference or an account reference. A service filter delivers webhooks for incidents belonging to that service, so one subscription per service or team keeps downstream receivers simple.
53. Name some event types.
Incident events such as incident.triggered, incident.acknowledged, incident.resolved, incident.priority_updated, incident.reassigned and incident.responder.added. There are also service events (service.created, service.updated, service.deleted) and workflow events (incident.workflow.started and incident.workflow.completed).
54. What does a v3 webhook payload contain?
One event object with id, event_type, resource_type, occurred_at, agent, client and data. The shape of data depends on the event type, so parse by event_type.
55. Can you add custom headers?
Yes. Custom headers are sent with each delivery, and their names must be unique within a subscription. Values are redacted in GET responses but not in the delivered webhooks, so use them for shared tokens your receiver checks, and treat the subscription as sensitive.
56. How should a receiver verify a webhook?
Check the X-PagerDuty-Signature header against an HMAC computed over the raw request body with the subscription's secret, and reject requests that do not match. Use the raw bytes, not a re-serialized body, and follow PagerDuty's Verifying Signatures page for the exact format.
57. What happened to webhooks v1 and v2?
PagerDuty says v2 reached end of support on 31 October 2022 with no end-of-life date set, so existing integrations keep working but get no features or fixes. V1 has been unsupported since 13 November 2021 and lost functionality on 31 October 2022. Migrate to v3.
Terraform and automation
58. Which Terraform resources manage integrations?
The PagerDuty provider includes resources such as pagerduty_service, pagerduty_service_integration, pagerduty_escalation_policy, pagerduty_schedule, pagerduty_event_orchestration and its global, router, service and unrouted variants, pagerduty_slack_connection, pagerduty_webhook_subscription and pagerduty_incident_workflow. The latest release at the time of writing is v3.37.0.
data "pagerduty_vendor" "datadog" {
name_regex = "Datadog"
}
resource "pagerduty_service_integration" "datadog" {
name = "Datadog"
service = pagerduty_service.checkout.id
vendor = data.pagerduty_vendor.datadog.id
}
59. Why manage Slack connections and webhooks as code?
They are easy to misconfigure by hand and hard to audit. Code review shows who connected which channel to which service, and a re-apply restores a deleted connection. Keep the API token in your secret store, not in the repository.
60. How do you onboard 30 services consistently?
Write a Terraform module that creates the service, its integration, an escalation policy reference, a Slack connection and an orchestration route, then call it per service with a few variables. Add a policy check in CI so no service ships without an owner team.
61. What triggers an Incident Workflow?
Conditional triggers (when an incident is created or matches conditions), manual triggers from the web app, mobile app, Slack or Microsoft Teams, and API triggers, which need a Manager, Admin or Account Owner role. Actions run in the order added, and a workflow left in draft does not run.
62. What can workflows do with other tools?
Typical actions include adding responders, creating a Slack channel, posting status updates and updating priority. PagerDuty also has a ServiceNow workflow action (Enterprise) that creates a ServiceNow incident. Use workflows for repeatable steps, not for decisions that need a person.
63. What are Automation Actions and runners?
Automation Actions let responders run diagnostic or remediation jobs from an incident with one click, tied to services. A runner executes the job and connects outward to PagerDuty, so no inbound firewall rule is needed. The feature is an add-on, so check plan availability.
64. How does Runbook Automation relate to Rundeck?
Rundeck is now offered as part of PagerDuty, and Runbook Automation is the hosted product built on it, with runners deployed behind firewalls or in VPCs. Newer documentation names two runner types, Process Automation and Runbook Automation, so confirm names in the current docs. For wider automation guardrails, read why you should automate incident response with runbooks.
Status pages and stakeholder communication
65. How do you create an external status page?
Go to Incidents, External Status Page, then New Status Page. You need the Manager, Global Admin or Account Owner base role. Choose the type, name, a Simple URL (a unique subdomain that cannot be changed after you save or publish), the business services to show, permissions and appearance.
66. Who can post to a status page, and can posts need approval?
The Create Post permission can be given to base roles or to specific teams. Approval is optional, and if you require it you can assign up to 15 reviewers. The user who creates a page becomes its Page Owner, and Account Owners can edit any page.
67. How do stakeholders get incident updates inside PagerDuty?
From the incident, the Status Updates tab lets you manage subscribers and send updates. If the option to post the resolution note as a status update is enabled, the note goes to all incident subscribers. For customer-facing messages use the external status page.
Troubleshooting scenarios
68. Alertmanager sent 200 pages in five minutes. What do you do?
Symptom: many incidents for one outage. Check the receiver's group_by and the grouping timers, add an inhibit rule so a cluster-wide alert mutes per-Pod warnings, and consider an Event Orchestration rule that suppresses known noise. After the incident, remove alerts nobody acted on.
69. A CloudWatch alarm fired, but there is no PagerDuty incident. Where do you look?
First confirm the SNS subscription status is Confirmed. Then check that raw message delivery is off, that the alarm's action points to the right topic, and that the endpoint is the integration URL for the service or orchestration. The integration needs the nested JSON Message, and without it no alert triggers.
70. The PagerDuty incident never resolves after the alert clears in Alertmanager. Why?
Check send_resolved on the receiver (default true), the group_interval delay, and that the resolve is sent with the same key as the trigger. A different routing key sends the resolve to a different place. Review Alertmanager logs for failed notifications.
71. Datadog shows a monitor as recovered, but the incident is open. What now?
Check that the monitor message includes the @pagerduty-ServiceName notification for recovery and that the integration is authorized. As a manual fix, comment @pagerduty-resolve on the original event stream post, and then find out why the automatic resolve did not arrive.
72. A webhook receiver gets no events. What do you check?
Verify the subscription exists, the event types include what you expect, the filter (service, team or account) matches the incident's service, and the destination URL is reachable from the internet. Then inspect the receiver's logs for rejected signatures or header mismatches.
73. A deploy at 10:05 was followed by an incident at 10:12. How does PagerDuty help?
If the pipeline sends change events to the service, the deploy shows on the service activity stream and the incident details page, so the responder can start with the change. Without them, someone reads pipeline logs while customers wait. Add the call to your deploy job and test it.
74. Engineers in Slack cannot acknowledge incidents. What is wrong?
Their Slack accounts are probably not linked to PagerDuty users, or their PagerDuty role does not allow the action. Link accounts, check roles, and ask a Slack admin to approve the app update if channel connections cannot be created.
75. A ServiceNow ticket and a PagerDuty incident disagree on priority. What do you check?
Priority syncs once priority mapping is configured in ServiceNow, and it syncs when an incident is triggered or edited. Compare the mapping table, then check whether the update was made in the system that owns priority. Decide which tool is the source of truth.
76. Customers ask whether the outage is over. How do you answer without paging the whole company?
Post an update on the external status page with the impact and next update time. Subscribers on the page get notified, and internal responders stay in the incident channel. Have the communications role own updates so the commander can keep diagnosing. More on roles in the purpose of SRE incident commanders.
How to prepare
- Build a lab with Prometheus, Alertmanager and a free PagerDuty trial or developer account. Page yourself from a real rule, then resolve it.
- Send a change event from a CI job and find it on the service activity stream.
- Create a v3 webhook subscription and print the payload in a small receiver.
- Write the Terraform for one service, integration, Slack connection and orchestration route.
- Explain a past outage as symptom, detection, routing, fix and prevention, in under two minutes.
For the monitoring half of the lab, our Prometheus and Grafana training in Pune covers Alertmanager configuration, and the DevOps training in Pune course covers the pipeline side. Both run in the classroom at Kharadi and live online across India. You may also want Prometheus and Grafana together and 10 incident response tools by lifecycle phase.
Frequently asked questions
How do you integrate Prometheus with PagerDuty?
Add a pagerduty_configs block to an Alertmanager receiver with the integration key as routing_key (Events API v2) or service_key (Prometheus integration), route the alerts to that receiver and reload Alertmanager. Resolve messages are sent by default.
Why does my CloudWatch alarm not create a PagerDuty incident?
Common causes are an unconfirmed SNS subscription, raw message delivery turned on, an alarm that does not publish to the topic, or a wrong integration URL. The integration needs the nested JSON message in the standard SNS format.
Do change events create incidents or send notifications?
No. A change event adds context to a service and its incidents but does not notify anyone. Send them from your pipeline after each production change.
Which webhook version should I use?
Use webhooks v3. PagerDuty states that v2 has been in end-of-support since 31 October 2022 and v1 is unsupported.
Can I manage PagerDuty integrations with Terraform?
Yes. The PagerDuty Terraform provider manages services, service integrations, escalation policies, schedules, Event Orchestration, Slack connections, webhook subscriptions and incident workflows.
How does the Slack integration help during an incident?
Responders can declare, acknowledge, escalate and resolve incidents, run workflows and post status updates from Slack. Dedicated incident channels can be created for each incident, and actions sync both ways.
What is the difference between a status page and incident updates?
An external status page is a public or private page for customers, with its own subscribers. Incident status updates go to the subscribers of one incident inside PagerDuty. Use the first for customers and the second for internal stakeholders.
Is this page the same as the main PagerDuty question set?
No. The main set covers the object model, Events API v2, escalation policies, schedules and limits. This page covers integrations with other tools and the troubleshooting that goes with them.
Official references
- PagerDuty: Events API v2 overview
- PagerDuty: send a change event
- PagerDuty: webhooks overview
- PagerDuty: Amazon CloudWatch integration guide
- PagerDuty: Slack integration guide
- PagerDuty: ServiceNow integration guide
- PagerDuty: Prometheus integration guide
- PagerDuty: Datadog integration guide
- PagerDuty: Event Orchestration
- PagerDuty: External Status Page
- Prometheus: Alertmanager configuration
- Microsoft: Send Azure service health alerts with PagerDuty using webhooks
- Grafana: PagerDuty contact point
- PagerDuty Terraform provider
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Angry
0
Sad
0
Wow
0