chore(ci): address two minor workflow issues (#438)

Quick fixes for two small issues:

- the `cancel-in-progress: true` on the automerge and auto-approve
workflows was sometimes (not consistently) causing _all_ runs of a
workflow to be canceled, instead of forcing only the most recent one to
run. This feels like a bug on GitHub's side but whatever, we should be
able to fix it by removing that setting.
- for whatever reason the alert-open-prs workflow has been failing a lot
this week with 503s/504s, which shows up as red on our dashboard. If
this workflow happens to fail once, it's not that big of a deal, so I
applied the same fix that we used for the lock workflow --
`continue-on-error: true` to have the workflow not show up as failed.

src/alert-open-prs.ts

@@ -71,6 +71,7 @@ export class AlertOpenPrs {
           env: { SLACK_WEBHOOK_URL: slackWebhookUrl },
         },
       ],
+      continueOnError: true, // this job fails sometimes due to Slack timeouts (503s and 504s) but it's really not a big deal if it does
     });
   }
 }

src/auto-approve.ts

@@ -23,7 +23,6 @@ export class AutoApprove {
 
     (workflow.concurrency as any) = {
       group: "${{ github.workflow }}-${{ github.head_ref }}",
-      cancelInProgress: true,
     };
 
     const maintainerStatuses = `fromJSON('["OWNER", "MEMBER", "COLLABORATOR"]')`;

src/automerge.ts

@@ -29,7 +29,6 @@ export class Automerge {
 
     (workflow.concurrency as any) = {
       group: "${{ github.workflow }}-${{ github.head_ref }}",
-      cancelInProgress: true,
     };
 
     const maintainerStatuses = `fromJSON('["OWNER", "MEMBER", "COLLABORATOR"]')`;