I've been running Sumo Logic for about six months to ingest logs from a hybrid Salesforce and custom app stack. For the last three weeks, the installed collector on our primary Ubuntu 20.04 server disconnects from the service at random intervals, sometimes multiple times a day.
The collector service remains running, but it stops sending data. The only fix is a manual restart. Our setup is standard:
* Local configuration file management (no cloud-based config).
* About 30 sources, mostly local file and script sources.
* We're well below any documented throughput limits.
Here's a sanitized snippet of the relevant `user.properties`:
```json
name=prod-collector-01
accessid=REDACTED
accesskey=REDACTED
sources=/opt/SumoCollector/sources.json
ephemeral=false
clobber=false
```
Before I spend two days building a watchdog script or opening a ticket, I want to see if this is a known issue with a documented fix. The Sumo status page has been clear.
Has anyone else reproduced this? Specifically:
* What OS and collector version are you on?
* Are you using any specific cloud provider or proxy configuration?
* Did you find any smoking gun in `collector.log` or `restart.log`?
I've checked the obvious: disk space, memory, network timeouts. Nothing spikes. The logs just show a normal shutdown sequence *after* the disconnect happens, which points to something upstream.
Show me the query.