Sentry and ELMAH Error Logging

VIDEO TRANSCRIPT | Recorded: 2022-01-27 | Verify against current system state

Abstract¶

Overview of Sentry and ELMAH error logging tools used for application monitoring at AANP. Demonstrates Sentry project setup, error analysis features including browser/device breakdowns, resolution workflows, and spike protection. Compares Sentry's modern interface with legacy ELMAH logging, discussing tradeoffs including Sentry's notification throttling behavior that can hide error floods.

Key Procedures¶

• Create new Sentry project for C# application
• Install Sentry NuGet package and configure in application
• View errors filtered by environment (production/staging)
• Analyze error frequency by time period
• Review browser, OS, and device breakdown for errors
• Assign errors to team members for investigation
• Mark errors as resolved after fix deployment
• Ignore errors temporarily with conditional thresholds
• Trash and discard future matching errors
• Share error details with team members
• Configure ELMAH for email and SQL Server logging
• Access ELMAH logs via elmah.axd endpoint

Notable Statements¶

• 0:00:05 "Sentry is one of the areas we use to track errors in our applications"
• 0:00:13 "This is a fairly recent addition. We just tried it out to see what it would be like"
• 0:00:29 "Any kind of exception that occurs, it really just handles all of them and pushes them into Sentry"
• 0:00:54 "The one site that we get the most errors on is CraftCMS, and that's a Linux site"
• 0:01:06 "The integration there is through CraftCMS just as a plug-in"
• 0:03:29 "You panic, because you see it's got 10,000 events. But the problem is, these are the total events over the lifetime"
• 0:04:56 "I could say, hey, Matt, take a look at this error, and it would send him a note"
• 0:05:28 "This error in particular, for the 10,000 times, it's happened. How many times has it happened in a specific browser?"
• 0:05:37 "47% of the time it was Chrome 81.0"
• 0:06:29 "When you click resolve, what happens is it starts sending you alerts again"
• 0:09:06 "We are subscribed to a hundred thousand limit and then 100,000 is per month"
• 0:09:40 "The only time I've ever seen it be hit was when there was a real problem on ww ampi.org where it couldn't redirect back to the SAML login piece"
• 0:10:27 "Spike protection... it's trying to make sure that we don't hit a 100k"
• 0:12:07 "This is like 23 $29 a month. The other ones were 100 $200 a month" DATED
• 0:12:45 "Sometimes it feels like it's hiding errors... Sentry will send you one notification and then it'll stop"
• 0:15:32 "These get logged here and they also get logged to a SQL Server database table"

Systems & Configurations¶

Systems Mentioned¶

• Sentry (error tracking SaaS)
• ELMAH (legacy error logging)
• CraftCMS (Linux-based CMS with plugin integration)
• GitHub (potential integration)
• React applications
• .NET Core applications
• SQL Server (ELMAH logging target)

Specific Configurations¶

Credentials/Access Mentioned¶

• Sentry project access for team members
• CraftCMS admin for plugin configuration
• confregqa.ampqa.com (QA environment URL)

Vendor Contacts Mentioned¶

• LogRocket mentioned as alternative (more expensive)

Errors & Troubleshooting¶

• Error: loginuser.php SAML login errors
• Cause: SAML authentication issues on CraftCMS
• Resolution: Ongoing investigation, using ignore feature temporarily

• Timestamp: 0:03:15

• Issue: Error count display confusion

• Cause: Total lifetime events shown vs recent events
• Resolution: Check 24-hour column for recent frequency

• Timestamp: 0:03:29

• Issue: Sentry hiding error floods

• Cause: Notification throttling after initial alert
• Resolution: Must actively monitor Sentry dashboard
• Timestamp: 0:12:45

Transcript Gaps & Quality Notes¶

• Screen sharing demonstration - UI details not captured in audio
• Reference to conference reg being down (seasonal)
• Team member Matt mentioned for error assignment
• Keith mentioned as knowledge transfer audience
• Some ELMAH URLs may not load (systems inactive)