From 4970b16d61e4377d483612b9a5849e82f438ac54 Mon Sep 17 00:00:00 2001 From: James Mills Date: Tue, 2 Feb 2021 12:07:53 +1000 Subject: [PATCH] Add better crawler metrics --- internal/crawler.go | 4 ++++ internal/server.go | 34 ++++++++++++++-------------------- 2 files changed, 18 insertions(+), 20 deletions(-) diff --git a/internal/crawler.go b/internal/crawler.go index 905ae28..ce8fc2c 100644 --- a/internal/crawler.go +++ b/internal/crawler.go @@ -48,6 +48,8 @@ func (c *crawler) loop() { log.Debugf("found %s", link) + metrics.Counter("crawler", "crawled").Inc() + url := NewURL(link) url.CrawledAt = time.Now() @@ -60,6 +62,8 @@ func (c *crawler) loop() { } else { if err := c.db.SetURL(hash, url); err != nil { log.WithError(err).Error("error recording url %s", link) + } else { + metrics.Counter("crawler", "scraped").Inc() } } } diff --git a/internal/server.go b/internal/server.go index 0390bbd..894e0ae 100644 --- a/internal/server.go +++ b/internal/server.go @@ -190,17 +190,22 @@ func (s *Server) setupMetrics() { return float64(s.db.LenTokens()) }, ) - - // feed cache sources - metrics.NewGauge( - "cache", "sources", - "Number of feed sources being fetched by the global feed cache", + metrics.NewGaugeFunc( + "db", "urls", + "Number of database /urls keys", + func() float64 { + return float64(s.db.URLCount()) + }, ) - // feed cache size - metrics.NewGauge( - "cache", "feeds", - "Number of unique feeds in the global feed cache", + // Crawler stats + metrics.NewCounter( + "crawler", "crawled", + "Number of links crawled by the crawler", + ) + metrics.NewCounter( + "crawler", "scraped", + "Number of links scraped by the crawler", ) // server info @@ -216,17 +221,6 @@ func (s *Server) setupMetrics() { "commit": spyda.Commit, }).Set(1) - // old avatars - metrics.NewCounter( - "media", "old_avatar", - "Count of old Avtar (PNG) conversions", - ) - // old media - metrics.NewCounter( - "media", "old_media", - "Count of old Media (PNG) served", - ) - s.AddRoute("GET", "/metrics", metrics.Handler()) }