Skip to content

Monitoring: adding long time storage for metrics #1656

@Tbaile

Description

@Tbaile

Firewall needs a long term storage for metrics, at the moment Netdata does the job but it has the following limitations:

  • only 6 hours
  • very old build
  • hard to customize

We landed on Victoria Metrics due to the nice space usage and easy replacement for Prometheus, which does the same as Netdata.
However a tool that scrapes the data from the firewall and send it to Victoria is needed, that's where Telegraf comes in.
From Telegraf the data to be sent to the Victoria Metrics is the following:

  • generic machine info (CPU, MEM, etc...)
  • MWAN status
  • services status
  • pings

Victoria can through vmalert mimic what AlertManager does, which is in firewall's case:

  • notify downtimes
  • notify alerts

Alerts is kept in storage by Victoria Metrics itself.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Projects

Status

In Progress 🛠

Relationships

None yet

Development

No branches or pull requests

Issue actions