I recently deployed this exporter, and it's functioning great, but I'm finding it's pretty slow in my environment. /metrics takes ~4.5 minutes, so I've configured my scrape timeout and interval appropriately, but it's still not ideal to have the requests running that long.
I'd love to have the ability to set up multiple scrapers, and have them hit the different collectors. I was kind of wondering whether the Prometheus library supported that natively, but it seems like it maybe doesn't.
My thinking is that we'd have, say:
/metrics -- fully backwards compatible
/metrics/client
/metrics/controller
/metrics/device
/metrics/port
Query string parameters would also work, but I think this is simpler to implement. But I'm really not familiar with the Prometheus libraries, so maybe there's a more clever solution.
This would also have the advantage that different metrics could be scraped at different intervals, if desired.
I recently deployed this exporter, and it's functioning great, but I'm finding it's pretty slow in my environment.
/metricstakes ~4.5 minutes, so I've configured my scrape timeout and interval appropriately, but it's still not ideal to have the requests running that long.I'd love to have the ability to set up multiple scrapers, and have them hit the different collectors. I was kind of wondering whether the Prometheus library supported that natively, but it seems like it maybe doesn't.
My thinking is that we'd have, say:
/metrics-- fully backwards compatible/metrics/client/metrics/controller/metrics/device/metrics/portQuery string parameters would also work, but I think this is simpler to implement. But I'm really not familiar with the Prometheus libraries, so maybe there's a more clever solution.
This would also have the advantage that different metrics could be scraped at different intervals, if desired.