Use Prometheus metrics prefix

Even though it is considered an anti-pattern to add a prefix with the name of the software or component to metrics (according to the official Prometheus documentation), I have decided to add a prefix. I’ve found that this makes it much easier to find relevant metrics. The main disadvantage of per-component prefixes queries become slightly more complex if you want to query the same metric (e.g. HTTP request duration) across multiple components. This isn’t super important in our case though, so I think the trade-off is acceptable.

ingestors/manager.py

@@ -25,24 +25,24 @@
 
 log = logging.getLogger(__name__)
 
-INGEST_SUCCEEDED = Counter(
-    "ingest_succeeded_total",
+INGESTIONS_SUCCEEDED = Counter(
+    "ingestfile_ingestions_succeeded_total",
     "Successful ingestions",
     ["ingestor"],
 )
-INGEST_FAILED = Counter(
-    "ingest_failed_total",
+INGESTIONS_FAILED = Counter(
+    "ingestfile_ingestions_failed_total",
     "Failed ingestions",
     ["ingestor"],
 )
-INGEST_DURATION = Histogram(
-    "ingest_duration_seconds",
+INGESTION_DURATION = Histogram(
+    "ingestfile_ingestion_duration_seconds",
     "Ingest duration by ingestor",
     ["ingestor"],
     # The bucket sizes are a rough guess right now, we might want to adjust
     # them later based on observed durations
     buckets=[
-        0.005
+        0.005,
         0.01,
         0.025,
         0.05,
@@ -57,8 +57,8 @@
         15 * 60,
     ],
 )
-INGEST_INGESTED_BYTES = Counter(
-    "ingest_ingested_bytes_total",
+INGESTED_BYTES = Counter(
+    "ingestfile_ingested_bytes_total",
     "Total number of bytes ingested",
     ["ingestor"],
 )
@@ -205,20 +205,20 @@ def ingest(self, file_path, entity, **kwargs):
             self.delegate(ingestor_class, file_path, entity)
             duration = max(0, default_timer() - start_time)
 
-            INGEST_SUCCEEDED.labels(ingestor_name).inc()
-            INGEST_DURATION.labels(ingestor_name).observe(duration)
+            INGESTIONS_SUCCEEDED.labels(ingestor_name).inc()
+            INGESTION_DURATION.labels(ingestor_name).observe(duration)
 
             if file_size is not None:
-                INGEST_INGESTED_BYTES.labels(ingestor_name).inc(file_size)
+                INGESTED_BYTES.labels(ingestor_name).inc(file_size)
 
             entity.set("processingStatus", self.STATUS_SUCCESS)
         except ProcessingException as pexc:
             log.exception("[%r] Failed to process: %s", entity, pexc)
 
             if ingestor_name:
-                INGEST_FAILED.labels(ingestor_name).inc()
+                INGESTIONS_FAILED.labels(ingestor_name).inc()
             else:
-                INGEST_FAILED.labels(None).inc()
+                INGESTIONS_FAILED.labels(None).inc()
 
             entity.set("processingError", stringify(pexc))
             capture_exception(pexc)

ingestors/support/convert.py

@@ -15,8 +15,8 @@
 TIMEOUT = 3600  # seconds
 CONVERT_RETRIES = 5
 
-INGEST_PDF_CACHE_ACCESSED = Counter(
-    "ingest_pdf_cache_accessed",
+PDF_CACHE_ACCESSED = Counter(
+    "ingestfile_pdf_cache_accessed",
     "Number of times the PDF cache has been accessed, by cache status",
     ["status"],
 )
@@ -32,12 +32,12 @@ def document_to_pdf(self, unique_tmpdir, file_path, entity):
             file_name = entity_filename(entity, extension="pdf")
             path = self.manager.load(pdf_hash, file_name=file_name)
             if path is not None:
-                INGEST_PDF_CACHE_ACCESSED.labels("hit").inc()
+                PDF_CACHE_ACCESSED.labels("hit").inc()
                 log.info("Using PDF cache: %s", file_name)
                 entity.set("pdfHash", pdf_hash)
                 return path
 
-        INGEST_PDF_CACHE_ACCESSED.labels("miss").inc()
+        PDF_CACHE_ACCESSED.labels("miss").inc()
         pdf_file = self._document_to_pdf(unique_tmpdir, file_path, entity)
         if pdf_file is not None:
             content_hash = self.manager.store(pdf_file)