restrict qa runs to a 'min_qa_crawler_image' if set in the chart (#1685)

- fixes #1684
- can be used to optionally restrict QA to only some crawls (eg. with
browsertrix-crawler>=1.0.0)
- enforce error on backend (return 400) and handle special error on the
frontend

backend/btrixcloud/crawls.py

@@ -3,6 +3,7 @@
 # pylint: disable=too-many-lines
 
 import json
+import os
 import re
 import contextlib
 import urllib.parse
@@ -64,6 +65,8 @@ class CrawlOps(BaseCrawlOps):
         self.colls.set_crawl_ops(self)
         self.event_webhook_ops.set_crawl_ops(self)
 
+        self.min_qa_crawler_image = os.environ.get("MIN_QA_CRAWLER_IMAGE")
+
     async def init_index(self):
         """init index for crawls db collection"""
         await self.crawls.create_index([("type", pymongo.HASHED)])
@@ -734,6 +737,14 @@ class CrawlOps(BaseCrawlOps):
         if crawl.state not in SUCCESSFUL_STATES:
             raise HTTPException(status_code=400, detail="crawl_did_not_succeed")
 
+        # if set, can only QA if crawl image is >= min_qa_crawler_image
+        if (
+            self.min_qa_crawler_image
+            and crawl.image
+            and crawl.image < self.min_qa_crawler_image
+        ):
+            raise HTTPException(status_code=400, detail="qa_not_supported_for_crawl")
+
         # can only run one QA at a time
         if crawl.qa:
             raise HTTPException(status_code=400, detail="qa_already_running")

chart/templates/configmap.yaml

@@ -52,6 +52,8 @@ data:
 
   CRAWLER_CHANNELS_JSON: "/ops-configs/crawler_channels.json"
 
+  MIN_QA_CRAWLER_IMAGE: "{{ .Values.min_qa_crawler_image }}"
+
 ---
 apiVersion: v1
 kind: ConfigMap

chart/values.yaml

@@ -209,6 +209,9 @@ crawler_pull_policy: "Always"
 
 crawler_namespace: "crawlers"
 
+# if set, will restrict QA to image names that are >= than this value
+# min_qa_crawler_image: ""
+
 # optional: enable to use a persist volume claim for all crawls
 # can be enabled to use a multi-write shared filesystem
 # crawler_pv_claim: "nfs-shared-crawls"

frontend/src/pages/org/archived-item-detail/archived-item-detail.ts

@@ -1217,10 +1217,16 @@ ${this.crawl?.description}
         icon: "check2-circle",
       });
     } catch (e: unknown) {
+      let message = msg("Sorry, couldn't start QA run at this time.");
+      if (e instanceof Error && e.message === "qa_not_supported_for_crawl") {
+        message = msg(
+          "Sorry, QA analysis is not supported for this crawl as it was run with an older crawler version. Please run a new crawl with the latest crawler and QA should be available.",
+        );
+      }
       console.debug(e);
 
       this.notify.toast({
-        message: msg("Sorry, couldn't start QA run at this time."),
+        message,
         variant: "danger",
         icon: "exclamation-octagon",
       });