A log ingestion pipeline almost never fails noisily. It goes quiet. An API token expires, a file changes format after an update, an overly broad ingest rule suddenly sucks in production debug logs. In all these cases nothing flashes: you discover the gap when you need the data, in the middle of an investigation, or at the end of the month when you read the consumption bill. Dynatrace log ingestion deserves the same monitoring as any other production service.
Dynatrace SaaS has improved visibility into this pipeline: an enriched ready-made dashboard, self-monitoring events emitted by the OneAgent log module, log module configuration grouped in the Settings app, and more complete sizing guides for the Environment ActiveGate. Taken separately, these are incremental improvements. Together, they finally let you treat log ingestion as a service in its own right, with its own monitoring.
Rather than reviewing the new features one by one, this article starts from the three most common failures and shows how to detect them, then suggests a checking routine to put in place.
Three ways to lose logs without knowing it
Silence
A source stops sending: unsupported encoding, a file compressed or renamed by rotation, a misconfigured time zone, a missing timestamp pattern. Historically, you had to inspect source by source to understand. The OneAgent log module now emits self-monitoring events (SFM) that describe source status and health signals. They are written to Grail, in dt.system.events, and each event type is documented with its meaning and a suggested fix.
Partial coverage
Here nothing has stopped: the source was simply never captured. A new host group, a recently added Kubernetes cluster, an application deployed without anyone creating the matching ingest rule. The difficulty is that this absence produces no signal. Only a coverage view makes it visible.
Excess
The opposite problem: a rule whose matcher is too permissive, a log level left on debug, an error loop flooding a file. Volume climbs, consumption follows, and the noise makes the useful logs harder to read.
The tools to monitor Dynatrace log ingestion
The Log ingest overview dashboard
This ready-made dashboard is the entry point. It now shows a health view of the ingestion components, for both the OneAgent log modules and the Environment ActiveGate, as well as a view of the top log producers. It highlights the components that need attention, which saves you building a query from scratch at the worst possible moment.
Log module configuration in Settings
Ingest rules, custom sources, masking and timestamp patterns are now brought together under Settings > Collect and capture > Log monitoring > Configure log module. The Sources tab opens on an autodiscovery summary: number of host groups, Kubernetes clusters, hosts and sources detected, with the percentage covered by a rule. Another welcome addition for anyone who has inherited a complex configuration: active rules and rules inherited from a higher level (environment or host group) are shown together, with a direct link to their definition. You can also ask Dynatrace Assist about coverage in natural language.
ActiveGate sizing guides
For logs sent by API (JSON or OTLP) to the Environment ActiveGate, the official documentation provides sizing guides covering containerised and host-based deployments, on Linux and Windows, with figures extended to large environments.
In practice
Step 1 — Make the dashboard a ritual
Open Log ingest overview in the Dashboards app and add it to the platform team’s favourites. A weekly glance at the component health section and the top producers is enough to spot most drifts before they become incidents.
Step 2 — Explore the self-monitoring events
When a component appears degraded, drill down to the events. An exploratory query in a Notebook gives you an overview; then refine the filters based on the fields you see in your environment:
fetch dt.system.events, from:now()-24h | filter event.provider == "Log Module" | summarize nb = count(), by:{event.type} | sort nb desc Then open a representative event: it shows the probable cause, for example an invalid time zone setting on a Kubernetes log module. The documentation page for the event type gives the fix to apply.
Step 3 — Measure coverage
In the Sources tab of the log module configuration, sort by coverage rate. A poorly covered host group is almost always the explanation for « we’re missing logs ». From the same row, examine the entity configuration and create the missing rule without leaving the page. While you are there, check the inherited rules: an exclusion defined at environment level can cancel out a local inclusion.
Step 4 — Track down excess volume
When the top producers view reveals an unexpected source, open it in the Logs app (log pattern analysis helps you understand what it produces) and look at the breakdown by severity level. A query like this gives the same information in a reproducible way:
fetch logs, from:now()-24h | summarize volume = count(), by:{log.source, loglevel} | sort volume desc | limit 15 If a production source is dominated by debug messages, look at recent changes to the ingest rules: a matcher meant for a development environment that also covers production is a classic.
Step 5 — Automate detection
Once the cause is fixed, create a log metric on the critical sources and attach a Dynatrace Intelligence alert on volume drops and spikes. Since self-monitoring events can also trigger Workflows, you can automatically open a ticket or notify the owning team as soon as a source goes into error.
Step 6 — Size the ActiveGate before a migration
If you are bringing logs into Dynatrace from another platform via OTLP, the calculation has two stages: first establish the base ActiveGate configuration for routing OneAgent traffic according to cluster size, then add the replicas needed for the expected OTLP log throughput, based on the per-replica capacity given in the documentation. Measure the source’s actual throughput before the cut-over rather than relying on an estimate.
Best practices
- Monitor the monitoring: treat ingestion health as a platform SLO, with an owner and alerts.
- Narrow the matchers: every ingest rule should explicitly target its scope (environment, namespace, host group).
- Review coverage after every infrastructure addition: a new cluster or a new host group means a new check of the Sources tab.
- Plan for token rotation: silent expiry is one of the most mundane causes of losing logs sent by API.
- Document inherited rules: the unified view helps, but an explicit name and description remain essential.
Availability
The Log ingest overview dashboard and the OneAgent log module self-monitoring events are available to Dynatrace SaaS customers using Grail and the latest version of the platform, with no extra configuration. Customers still on Log Monitoring Classic need to enable these events explicitly.
Frequently asked questions
Why does a log source stop arriving without any alert?
The causes are often mundane: an expired API token, a file renamed or compressed by rotation, an unsupported encoding, a misconfigured time zone or timestamp pattern. The self-monitoring events from the OneAgent log module, stored in dt.system.events, describe these situations and point to the fix.
How do you check that every source is covered?
In the Sources tab of the log module configuration, sort by coverage rate: a poorly covered host group almost always explains missing logs. Also check inherited rules, which can cancel out a local inclusion.
Are these improvements available in Log Monitoring Classic?
The dashboard and self-monitoring events are available with no configuration for SaaS customers on Grail and the latest platform version. In Log Monitoring Classic, the events must be enabled explicitly.
Conclusion
Most log losses are not spectacular outages but quiet drifts. By combining the ingestion dashboard, the self-monitoring events stored in Grail, the log module coverage view and volume alerts, a platform team can move from a reactive mode, where gaps are discovered mid-crisis, to one where ingestion is monitored and fixed like any other service.
Further reading: Modern log management just got simpler: Resilience and self-service upgrades for log ingestion (Dynatrace blog).
span.observer is an independent community blog, not affiliated with Dynatrace LLC. Dynatrace is a trademark of Dynatrace LLC. SaaS features evolve quickly: check their availability in your environment and in the official documentation.
Originally published on span.observer. French version: https://span.observer/fr/ingestion-logs-resilience-self-service/.

