Ingester queue is getting too deep

[tales-aparecida]

Reported at https://discord.com/channels/1245820301053530313/1301896040349433957/1441194713188274176

Possible actionables

• Check where is the bottleneck (Python, PostgreSQL, or origins re-submitting data)
• If it is Python: Check if we could deploy multiple ingesters in parallel
• If it is PostgreSQL: Check if settings or indexes that would improve the consumption
• If it is origins: Check which origin is causing the bursts
• Rate-limit submissions
• Rewrite ingester in Rust, if we determine Python is the culprit