Observability
Every integration sends its logs, metrics and traces through OpenTelemetry, each tagged with the integration, version, project, environment and agent it came from. Weik.io checks them every 15 seconds and raises a problem when an integration fails, restarts in a loop, grows its memory day after day or stops handling messages. You read all of it in the admin UI or with the weikio CLI, and you can forward it to Azure Monitor, Splunk or any OTLP receiver.
$ weikio logs finance/erp-invoice-export --env production -f
$ weikio metrics finance/erp-invoice-export --env production
$ weikio traces finance/erp-invoice-export --env production --errors
$ weikio trace 4bf92f3577b34da6a3ce929d0e0e4736
$ weikio problems --env productionIn the admin UI
Problems Weik.io raises
Weik.io raises these problems, among others, and each one shows on Home, in Observability, on the record's page and in weikio problems:
- An integration that fails at least 5 messages and 5 % of its messages in 15 minutes
- An integration that restarts in a loop
- Memory that grows day after day
- An integration that stopped handling messages
- An API failing on the server's side
- A failed transfer run
- Storage filling up
- Telemetry not arriving at Weik.io
- A backup overdue
- Two-factor sign-in switched off for platform administrators
Your own numbers
Counters and timers written with Camel's micrometer: and opentelemetry-metrics: components show up next to the platform's own numbers, with the same integration, version and environment on them. The video Custom metrics adds a counter to an integration and finds it in the admin UI.
- route:
from:
uri: direct:export-invoice
steps:
- to: "{{erp-url}}/api/invoices"
- to: micrometer:counter:invoices_processedHow long it's kept
| Signal | Kept |
|---|---|
| Logs | 30 days |
| Metrics | 90 days |
| Traces | 7 days |
| Audit log | 400 days |
An admin sets each signal's retention in the installation settings, with weikio settings set --logs-retention, --metrics-retention and --traces-retention, and the log and trace stores keep fewer days when they reach their disk cap.
Your own tools
A platform admin applies a telemetry-forwarding document, and an OpenTelemetry Collector in Weik.io forwards logs, metrics, traces and the audit trail to your own systems. The docs have complete examples for Azure Monitor, Splunk HEC and an OTLP SIEM, and Settings › Telemetry forwarding and weikio forwarding status show that the records arrive. Every problem Weik.io raises also writes a log line and counts in a metric, and forwarding brings both to your own alerting.
Every process of the platform also serves its metrics in Prometheus format, so your own monitoring can watch Weik.io and its agents.
When the link is down
An agent keeps up to 4 GiB of telemetry on its own disk while Weik.io is out of reach, or a tenth of its free disk when that is less. When the link is back it sends the records with their original times, and the live tail shows those late lines too.
Open the live demo
The live demo has a failing integration on purpose, so every screen on this page has something to show.