Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteFor a new Grafana-based logging deployment, use Grafana Alloy—not Promtail. Grafana declared Promtail end of life on March 2, 2026; it is no longer maintained or updated. Alloy can collect application logs from Kafka topics, collect Kubernetes pod logs directly, or do both, and forward entries to Loki for exploration in Grafana. The right collection path depends on where your logs already are: Kubernetes logs do not need to pass through Kafka unless your architecture calls for it.
How centralized logging fits together
A typical flow has three parts: a collector reads logs from their source, Loki stores and organizes the entries, and Grafana connects to Loki so people can search and explore them with LogQL. In this guide, Alloy is the current Grafana collector for both Kafka and Kubernetes sources.
There are two distinct collection paths. If application logs already arrive on Kafka, Alloy can consume configured topics and forward their records to Loki. If the logs are Kubernetes pod output, Alloy can discover pods and collect their logs directly. A deployment can use either path or both; the Kafka tutorial’s example is illustrative and is not presented as the usual way every application should be wired.
Choose where Alloy should collect logs
| Collection path | Use it when | What Alloy reads | Operational consideration |
|---|---|---|---|
| Kafka topics | Application log records are already published to Kafka, or Kafka is the intended handoff point. | Configured brokers and topics; a consumer group is used to consume the configured stream. | The team responsible for the Kafka-to-Loki path must manage topic access, consumer configuration, and any relabeling. |
| Kubernetes pod logs | You want to collect container output from Kubernetes without routing it through Kafka. | Pods discovered in Kubernetes and their logs, using Alloy’s Kubernetes discovery and Loki source components. | The team operating the Kubernetes collector owns discovery and collection configuration. |
Grafana’s Alloy Kafka example demonstrates a loki topic containing structured JSON and an otlp topic containing serialized OpenTelemetry log data. Treat those as examples of supported input patterns, not required topic names or formats. Kafka metadata and relabeling requirements also depend on your setup; decide which useful source dimensions should become Loki labels before adopting a configuration.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
- Metamorphosis: Franz Kafka (Little Clothbound Classics)
Deploy Loki for your environment
Grafana documents several self-managed installation routes, including Helm, Tanka, Docker or Compose, local execution, and building from source. Its installation documentation recommends Helm as one route. Pick an installation method and topology that fit the cluster and operating requirements rather than assuming an introductory setup is production-ready.
Grafana’s getting-started material demonstrates Loki in monolithic, single-binary mode on Kubernetes. That is an introductory path, not a universal production architecture. Storage choice, log volume, retention, availability requirements, and operational capacity affect the appropriate deployment; no single topology or sizing target is established for every cluster. For object storage, consult Grafana’s documented authentication requirements for the storage provider you use.
If you prefer a managed service, Grafana Cloud Logs is an alternative to operating Loki yourself. Compare operational responsibility, data residency, retention needs, integration constraints, and the current plan terms for your requirements; the available information does not establish a general cost winner.
Configure Alloy and send logs to Loki
For Kafka application logs
Configure Alloy’s Kafka source with the brokers and topics it should consume, then connect that source to Loki’s write component. Select a consumer group appropriate to your deployment, and add relabel rules only where needed to map useful source metadata into stream labels. The exact broker access, topic names, record formats, and rules are environment-specific; the official example’s topic names should not be copied as assumptions.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
For Kubernetes pod logs
Use Alloy’s Kubernetes discovery and Loki Kubernetes log collection components to discover and read pod logs, then send the resulting entries to Loki. Grafana’s getting-started example uses this direct collection path. It can coexist with Kafka collection when different log sources require different routes.
Connect Grafana to Loki
Configure Loki as a data source in Grafana. In Explore, choose a time range and use LogQL to narrow the logs to the streams and entries you need. Keep the flow clear during setup: first verify that Alloy is sending the intended records to Loki, then verify that Grafana’s data source can query them.
Design labels for useful searches
Loki indexes labels that identify log streams; it does not index the full contents of every log line in the same way a full-text index would. After narrowing to streams with labels, the log line itself remains searchable. This makes label choice part of the query workflow: labels should help select a useful subset without turning every changing or high-cardinality value into a stream dimension.
Grafana’s examples use source dimensions such as region, cluster, or environment, and the Kubernetes sample includes container and pod labels. Choose a small set that helps your operators distinguish meaningful sources, then use LogQL to search within the selected streams and time range. Avoid treating labels as a substitute for the message contents or assuming every JSON field is automatically indexed.
Secure Loki before exposing it
Loki does not include an authentication layer. Grafana’s installation documentation recommends putting an authenticating reverse proxy or equivalent protection in front of Loki services as appropriate to the deployment. Do not expose an unauthenticated Loki endpoint simply because the collector and Grafana data source are configured to communicate with it; determine who can reach each service and how requests are authenticated before making it accessible.
Migrate existing Promtail configurations to Alloy
Promtail reached end of life on March 2, 2026. Grafana states that commercial support has ended, no further support or updates will be provided, and future feature development is in Alloy. Existing users are directed to migrate to Alloy or another supported client.
- Run Alloy’s Promtail conversion command to generate a starting Alloy configuration from the existing Promtail configuration.
- Review conversion diagnostics and resolve issues rather than assuming a successful conversion guarantees identical behavior.
- Check configuration differences, including the positions-file location and monitoring metric names, and update any dependent operational tooling.
- Test the converted setup against representative log sources and queries before production cutover. Grafana warns that bypassing conversion errors may result in behavior that does not match the original configuration.
A converted file is a migration aid, not proof of production equivalence. Compare the actual logs and labels reaching Loki, and validate the dashboards and alerts that depend on them.
Decide who operates the logging path
The collection choice and the Loki hosting choice are separate decisions. Kafka collection is useful when Kafka is already the log handoff or when that stream is part of the desired architecture; direct pod collection avoids making Kafka a mandatory intermediary for Kubernetes output. Separately, self-managed Loki gives your organization responsibility for deployment, storage, security, and retention operations, while Grafana Cloud Logs is a managed-service path. Workload volume, retention, security, residency, and operational requirements determine which combination fits; there is no workload-specific performance, sizing, or cost comparison established here.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →
