10 Real-Time Logging Tools for DevOps Teams (2027)
Compare 10 real-time logging tools: ELK, OpenSearch, Loki, Graylog, VictoriaLogs, SigNoz, Splunk, Datadog, Sumo Logic, Papertrail. Live tail and licenses.
Real-time logging tools collect logs from servers, containers and cloud services, make them searchable within seconds, and alert when a pattern appears. The Elastic Stack, OpenSearch, Grafana Loki, Graylog, VictoriaLogs and SigNoz can be self-hosted. Splunk, Datadog, Sumo Logic and Papertrail are hosted platforms with live tail views for watching logs as they arrive.
Logs are one of the three observability signals. For metrics and traces, and for how logging fits a wider stack, read our pillar post on the top observability tools for DevOps monitoring. This page focuses on one question: how fast each logging tool gets a log line from your application to a search result or an alert, and what that costs to run.
Licenses, ownership, commands and query syntax were checked against each vendor's documentation in November 2026.
Key takeaways
- "Real time" in logging means two things: a live tail view of incoming logs, and a short delay between ingestion and searchability. Elasticsearch, for example, refreshes actively searched indices every second by default.
- The choice is mainly about indexing. Elasticsearch, OpenSearch and Graylog index log content for fast full-text search. Loki indexes only labels, which is cheaper to store but slower for broad text searches. VictoriaLogs and SigNoz (on ClickHouse) use columnar storage.
- Licenses vary: OpenSearch and VictoriaLogs are Apache-2.0, Loki is AGPLv3, Graylog Open is SSPL, Elasticsearch offers AGPLv3, SSPL or Elastic License 2.0, and SigNoz is MIT outside its enterprise directory.
- Promtail reached end of life on 2 March 2026. Ship logs to Loki with Grafana Alloy, Fluent Bit or an OpenTelemetry Collector instead.
- Splunk is now part of Cisco, and Papertrail is sold as SolarWinds Papertrail.
- Use a separate, lightweight shipper (Fluent Bit, Vector, Alloy or the OpenTelemetry Collector) so you can change the backend without touching every node.
How a real-time logging pipeline works
Every tool on this list sits somewhere in the same pipeline. Knowing the stages helps you see where delay creeps in.
- Collect. An agent on each node or a sidecar reads log files, journald or container output. In Kubernetes this is usually a DaemonSet reading
/var/log/containers. - Process. The agent parses lines into fields, adds metadata such as pod and namespace, drops noise and masks sensitive values.
- Buffer. High-volume setups put Kafka or the agent's own disk buffer between collection and storage so a slow backend does not lose data.
- Store and index. The backend writes logs and builds whatever index it uses.
- Query and alert. Users search, live tail and build dashboards. Alert rules run queries on a schedule and notify when a threshold is crossed.
Most of the delay comes from agent flush intervals, buffering and the backend's refresh cycle. A well-tuned pipeline usually makes logs searchable within seconds. If you are just starting, read why teams should adopt centralized logging early.
Real-time logging tools comparison
| Tool | Model | License | Indexing | Query language | Live tail |
|---|---|---|---|---|---|
| Elastic Stack | Self-hosted or Elastic Cloud | AGPLv3, SSPL or ELv2 | Full-text inverted index | KQL, ES|QL, Query DSL | Kibana Discover with auto-refresh |
| OpenSearch | Self-hosted or managed | Apache-2.0 | Full-text inverted index | DQL, PPL, SQL, Query DSL | Dashboards with auto-refresh |
| Grafana Loki | Self-hosted or Grafana Cloud | AGPLv3 | Labels only, content in chunks | LogQL | Grafana Explore Live, logcli --tail |
| Graylog | Self-hosted or cloud | Open edition SSPL; paid editions | Full-text (OpenSearch-based storage) | Graylog search syntax | Streams with real-time alerts |
| VictoriaLogs | Self-hosted | Apache-2.0 | Columnar with stream fields | LogsQL | /select/logsql/tail |
| SigNoz | Self-hosted or cloud | MIT (ee directory separate) | Columnar (ClickHouse) | Query builder, ClickHouse SQL | Live View in the Logs Explorer |
| Splunk | Splunk Cloud Platform or Enterprise | Commercial | Index time plus search-time fields | SPL | Real-time searches |
| Datadog Log Management | SaaS | Commercial | Indexed and non-indexed tiers | Datadog log search | Live Tail, including non-indexed logs |
| Sumo Logic | SaaS | Commercial | Platform managed | Sumo Logic query language | Live Tail per source or collector |
| SolarWinds Papertrail | SaaS | Commercial | Platform managed | Simple search | Live tail is the core feature |
Self-hosted and open source backends
1. Elastic Stack (ELK)
Elasticsearch stores and indexes logs, Kibana searches and visualises them, and Elastic Agent, Beats or Logstash ship and transform data. Because every word of every log line is indexed, ad hoc full-text searches are fast, and ES|QL adds piped queries for analytics. Elasticsearch is near real time: it refreshes indices every second by default, but only indices that received a search in the last 30 seconds, which saves resources on idle indices.
Elasticsearch and Kibana source code is available under AGPLv3, SSPL or Elastic License 2.0; Elastic added the AGPL option in September 2024. Full-text indexing needs more disk and memory than label-based stores, so plan index lifecycle policies and retention from day one. Our post on the ELK Stack in centralized logging covers the architecture in more detail.
2. OpenSearch
OpenSearch is the Apache-2.0 fork of Elasticsearch and Kibana, now run under the Linux Foundation through the OpenSearch Software Foundation. It provides the same style of full-text log search with OpenSearch Dashboards, and an Alerting plugin with per query, per bucket, per document and composite monitors that send to notification channels. Amazon OpenSearch Service and other providers offer it as a managed service.
Choose OpenSearch over Elasticsearch when you want a permissive license or use a managed OpenSearch offering. Expect API differences in features added after the 2021 fork.
3. Grafana Loki
Loki indexes only a small set of labels, such as namespace, app and pod, and stores the log text in compressed chunks on object storage. This makes it much cheaper to run at volume than a full-text index, provided you keep label cardinality low. LogQL filters by labels first and then scans lines:
{namespace="shop", app="checkout"} |= "error" != "healthcheck"
sum by (pod) (rate({namespace="shop", app="checkout"} |= "error" [5m]))
For live viewing, use the Live mode in Grafana Explore or the CLI:
export LOKI_ADDR=http://loki.monitoring:3100
logcli query --tail '{namespace="shop", app="checkout"}'
Alerting works through Loki's ruler, which evaluates LogQL rules in the Prometheus rule format:
groups:
- name: checkout-logs
rules:
- alert: CheckoutErrorSpike
expr: sum(rate({namespace="shop", app="checkout"} |= "error" [5m])) > 5
for: 5m
labels:
severity: page
annotations:
summary: "Checkout is logging more than 5 errors per second"
Loki is AGPLv3. Promtail, its original agent, reached end of life on 2 March 2026; Grafana Alloy replaces it and can convert Promtail configuration. Our pillar post compares Loki and Elasticsearch in more depth.
4. Graylog
Graylog is a log management platform that stores messages in an OpenSearch-based backend (the server code also ships Elasticsearch 7 storage support). Incoming messages are routed into streams, processed by pipeline rules, and can trigger real-time alerts. Editions include Graylog Open, Graylog Enterprise (with a data lake and AI summaries) and Graylog Security, a SIEM. Graylog Open is released under the Server Side Public License (SSPL) version 1, which is source available rather than OSI open source.
Graylog suits teams that want a ready-made log UI with role-based access and alerting, without assembling Elasticsearch or OpenSearch, Kibana and alerting themselves.
5. VictoriaLogs
VictoriaLogs, from the VictoriaMetrics team, is an Apache-2.0 log database with single-node and cluster versions. It accepts logs from common collectors and protocols, including Loki, OpenTelemetry, syslog, journald and Datadog formats, and has a Grafana data source plugin. Its query language is LogsQL, and live tailing is built in:
# Live tail error logs (default port 9428); -N stops curl buffering
curl -N http://localhost:9428/select/logsql/tail -d 'query=error'
# Count recent logs by level
curl http://localhost:9428/select/logsql/query \
-d 'query=_time:5m log.level:* | stats by (log.level) count() matching_logs'
By default the tail endpoint checks for new logs every second and delays output by 5 seconds so late-arriving logs are not missed. VictoriaLogs is a good fit for teams already running VictoriaMetrics, or anyone who wants full-text search with lower resource use than Elasticsearch.
6. SigNoz
SigNoz is an OpenTelemetry-native observability platform that stores logs, metrics and traces in one columnar database (ClickHouse), so you can jump from a log line to the trace that produced it. Its Logs Explorer has a Live View that streams new entries matching the current query. The Community edition self-hosts on Docker, Kubernetes or Linux; SigNoz Cloud and an enterprise self-hosted edition are paid. The repository is MIT licensed except the ee/ directory, which has its own license.
Choose SigNoz if you are standardising on OpenTelemetry and want logs correlated with traces in one open source tool.
Hosted platforms
7. Splunk
Splunk is the long-standing enterprise choice for log analytics and security operations, sold as Splunk Enterprise (on-premises, private cloud or hybrid) and Splunk Cloud Platform. Its Search Processing Language (SPL) is powerful for ad hoc investigation and reporting, and many security teams run their SIEM on Splunk. Splunk is now part of Cisco, and its observability products are sold separately as Splunk Observability Cloud.
Splunk is usually chosen for compliance, security analytics and organisations that already have Splunk skills. Ingest-based licensing can be expensive at high volume, so filter and route data before it reaches Splunk.
8. Datadog Log Management
Datadog collects logs through its agent and links them to metrics, traces and real user monitoring in one SaaS platform. Live Tail shows log events in near real time as they arrive, for both indexed and non-indexed logs, after Log Pipelines have parsed and enriched them. When a query matches many logs, Live Tail samples the output, so narrow the filter to see every event. You can index only the logs you search often and keep the rest in cheaper storage.
Datadog fits teams that already use it for infrastructure or APM and want logs in the same place. Watch indexing volume, which drives cost.
9. Sumo Logic
Sumo Logic is a cloud log analytics and SIEM platform that also offers data pipelines and AI agents for triage. Its Live Tail gives a real-time feed of messages from a chosen source or collector, with keyword filters and highlighting, and a CLI version. It offers a free trial rather than a permanent free tier on its homepage.
Teams pick Sumo Logic when they want one hosted service for operational logs and security analytics, especially on AWS and Kubernetes.
10. SolarWinds Papertrail
Papertrail is a simple hosted log service built around live tailing. Logs from many sources land in one viewer where you can search, tail and turn a saved search into an alert with instant or daily notifications. It offers a 30-day free trial. It is a practical choice for small teams and side projects that need centralised logs quickly without running anything.
Shipping logs: the collector layer
Pick the collector separately from the backend. These four tools can send to most of the backends above:
| Collector | License | Notes |
|---|---|---|
| Fluent Bit | Apache-2.0 | Lightweight agent; CNCF graduated sub-project of Fluentd; YAML or classic configuration |
| Vector | MPL-2.0 | Written in Rust and maintained by Datadog; transforms with Vector Remap Language |
| Grafana Alloy | Apache-2.0 | Grafana's OpenTelemetry Collector distribution; replaces Promtail and Grafana Agent |
| OpenTelemetry Collector | Apache-2.0 | Vendor-neutral; one pipeline for logs, metrics and traces |
A Fluent Bit configuration that tails container logs in Kubernetes, adds pod metadata and sends to Loki looks like this:
service:
flush: 1
log_level: info
pipeline:
inputs:
- name: tail
path: /var/log/containers/*.log
multiline.parser: docker, cri
tag: kube.*
filters:
- name: kubernetes
match: kube.*
outputs:
- name: loki
match: kube.*
host: loki.monitoring.svc
port: 3100
labels: job=fluent-bit
auto_kubernetes_labels: on
Be careful with auto_kubernetes_labels: every pod label becomes a Loki stream label, and high-cardinality labels slow Loki down. For a wider list of Kubernetes logging options, see Kubernetes logging tools for application insights and the Kubernetes logging mistakes beginners make.
Getting real-time value from logs
- Log in structured JSON. Parsed fields such as level, service, trace ID and user ID make searches and alerts precise. Parsing free text at query time is slow.
- Add the trace ID. With OpenTelemetry, every log line can carry the trace and span ID, so you can move from an error to the full request path.
- Alert on rates, not single lines. One error line is noise. A rise in errors per second for one service is a signal. Compare against metrics alerts in Prometheus, described in using Prometheus and Grafana together.
- Drop and sample at the edge. Debug logs and health checks can be filtered in the collector before they cost money in storage.
- Mask secrets and personal data before shipping. Once a password or token is indexed, it is copied into backups and replicas.
- Set retention by value. Keep hot, searchable logs for days or weeks and move older logs to object storage for audits.
Which logging tool should you choose?
| Situation | Good starting point |
|---|---|
| Kubernetes, cost-sensitive, already on Grafana | Grafana Loki with Alloy or Fluent Bit |
| Heavy full-text search or security analytics, self-hosted | Elastic Stack or OpenSearch |
| Permissive license required | OpenSearch or VictoriaLogs |
| Ready-made log UI with alerting | Graylog |
| OpenTelemetry everywhere, logs linked to traces | SigNoz |
| Enterprise compliance and SIEM | Splunk or Sumo Logic |
| Already using Datadog for APM | Datadog Log Management |
| Small team, minimal setup | SolarWinds Papertrail |
Other tools you will meet include Logz.io, Better Stack, New Relic logs, OpenObserve, Google Cloud Logging, Amazon CloudWatch Logs and Azure Monitor Logs. For a feature-by-feature view of more tools, see our 20 DevOps logging tools compared.
Learn logging and monitoring hands-on
Our ELK Stack training covers Elasticsearch, Logstash, Beats and Kibana with real log pipelines, and our Prometheus and Grafana training covers metrics, dashboards and alerting. Both run in Kharadi, Pune and live online across India.
FAQs
What is a real-time logging tool?
It is a system that collects logs as they are written, makes them searchable within seconds and lets you watch or alert on them live. Most tools offer a live tail view, near real-time indexing and alert rules that run log queries on a schedule.
Is the ELK Stack still free?
The source code of Elasticsearch and Kibana is available under AGPLv3, SSPL or Elastic License 2.0, so you can self-host it without paying. Some features require a paid subscription. OpenSearch is the Apache-2.0 alternative if you need a permissive license.
Should I use Loki or Elasticsearch for logs?
Use Loki when you mostly filter by labels such as namespace and app and want low storage cost. Use Elasticsearch or OpenSearch when you need fast full-text search across log content and complex analytics, and can afford the extra indexing resources.
What replaced Promtail?
Grafana Alloy. Promtail reached end of life on 2 March 2026, and Alloy can convert Promtail configuration. Fluent Bit and the OpenTelemetry Collector can also send logs to Loki.
What is the difference between Fluent Bit and Fluentd?
Both are CNCF graduated log processors from the same project. Fluent Bit is lightweight and suited to running on every node as an agent. Fluentd has a large plugin ecosystem and is often used as a central aggregator. Many setups run Fluent Bit on nodes and forward to a backend directly.
Which logging tools have live tail?
Datadog Live Tail, Sumo Logic Live Tail and Papertrail are built around it. Loki supports live tailing in Grafana Explore and with logcli query --tail, and VictoriaLogs has a /select/logsql/tail endpoint. Kibana and OpenSearch Dashboards use auto-refresh on recent time ranges.
Is Graylog open source?
Graylog Open is free to use, but it is released under the Server Side Public License version 1, which the Open Source Initiative does not approve as open source. Graylog Enterprise and Graylog Security are paid editions.
How long should I keep logs?
Keep searchable logs for as long as you actively investigate incidents, often 7 to 30 days, and archive older logs to cheaper object storage. Compliance rules in your industry may require longer retention, so check them before setting policies.
References
- Elastic: Near real-time search
- Elastic licensing FAQ
- OpenSearch repository
- OpenSearch Software Foundation
- OpenSearch Alerting
- Grafana Loki: LogQL
- Grafana Loki: LogCLI
- Migrate from Promtail to Grafana Alloy
- Graylog server repository
- Graylog products
- VictoriaLogs documentation
- VictoriaLogs: Querying and live tailing
- SigNoz repository
- Splunk
- Datadog: Live Tail
- Sumo Logic: Live Tail
- SolarWinds Papertrail
- Fluent Bit manual
- Fluent Bit: Loki output
- Vector repository
- Grafana Alloy repository
- SigNoz Logs Explorer
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Angry
0
Sad
0
Wow
0