You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current implementation (as of 2.4.3) of the homeautomation metrics makes several calls to the box per device found, which adds up to 20-30s per scrape, which may hit the Prometheus scape_timeout.
The code needs to be optimized, as it currently is pretty naive. AHA devices should prpbably only be scraped using the HTTP API and not a combination of TR-064 and HTTP to reduce the number of calls.
Just to mention here, the FB I'm testing against is a 7530 on a 1 Gbit Ethernet connection. I have very little info to go on what the CPU difference between models is.
This was not to say, that Fritz boxes are slow, just that they maybe too slow, if the exporter does not try to be "smarter" when reading the stats. ;-)
The current implementation (as of 2.4.3) of the homeautomation metrics makes several calls to the box per device found, which adds up to 20-30s per scrape, which may hit the Prometheus scape_timeout.
The code needs to be optimized, as it currently is pretty naive. AHA devices should prpbably only be scraped using the HTTP API and not a combination of TR-064 and HTTP to reduce the number of calls.
Also see the discussion in #303.
The text was updated successfully, but these errors were encountered: