feat: add monitoring metrics to process_pending container

[gustavobtflores]

Description

Add Prometheus metrics and a Grafana dashboard for the pending aggregations processor (process_pending_aggregations)

Changes

• Instrument process_pending_aggregations with Prometheus metrics (aggregation_records_written_total by table, aggregation_last_success_timestamp_seconds).
• Expose the processor metrics endpoint via Docker Compose (8003:8001) and add a Prometheus scrape job for it.
• Add the monitoring/aggregation_process.json Grafana dashboard and update monitoring docs with import steps.

How to test

1. docker compose -f docker-compose.monitoring.yml up -d
2. docker compose --profile with_commands up -d pending_aggregations_processor
3. In Prometheus (http://localhost:9090/targets), confirm kernelci-pending-aggregations-processor is UP, then import monitoring/aggregation_process.json in Grafana to verify panels populate.

[AmadeusK525]

Saving a timestamp is really weird imo, doesn't really fit with a timeseries record. Instead, I would just add a Counter that counts how many successful writes there were (isn't that what AGGREGATION_RECORDS_WRITTEN does?). We can know the approximate "last successful timestamp" simply by looking at the counter, if it stops going up then we have a problem.

[AmadeusK525]

High resolution here, I think 10s should be more than enough, no? How often does the cron job for processing pending even run?

[gustavobtflores]

Right now it seems to be configured to run each 1s, anyway we can increase the resolution here for sure

[MarceloRobert]

is there a need for a default here seen that you already have a default in the environment section?

[MarceloRobert]

LGTM, it's working well