Set up analytics and monitoring

See what subscribers do in the portal, what breaks for them, and what subscriptions earn, in tools you already use.

10 min

Three optional connections on Settings > Integrations watch instead of send: Triple Whale receives subscription events for attribution, PostHog shows portal usage, and Sentry reports portal errors. Connect each with a key from your own account; the senders (Klaviyo, SendGrid, Twilio) have their own page.

1. Forward subscription events to Triple Whale

   • Open Settings > Integrations, find Triple Whale under Analytics providers, and click Connect.
   • Paste an API key created in Triple Whale under Settings > API Keys with the data-in scope, then click Connect.
   • Every subscription change then flows by itself: created, active, paused, cancelled, with the cancellation reason.
   • No test button; proof is data arriving after the next subscription event. Disconnecting stops it.

   Checkpoint: the Triple Whale card reads Connected.

2. Watch portal usage in PostHog

   • Find PostHog under Analytics providers and click Connect.
   • Project API key is the phc_ key from PostHog under Settings > Project. API host stays blank for PostHog Cloud US; https://eu.i.posthog.com for EU Cloud, or your self-hosted address.
   • Click Connect, then Test connection: it captures an OpenSubs Test Event you can watch arrive in PostHog.
   • Portal visits are tracked like any site, subscription lifecycle events are captured server-side, and after login both land on one PostHog person, keyed by email.
   • Only the portal is tracked, never your admin; it loads its own PostHog instance, separate from any in your storefront theme.

   

   Checkpoint: the test event shows in PostHog.

   Events arrive scrubbed. Login codes and click-to-act links are stripped before events leave for PostHog.

3. Send portal errors to Sentry

   • Find Sentry under Monitoring and click Connect.
   • DSN comes from your Sentry project under Settings > Client Keys (DSN). Environment is optional and tags every report; production is the usual choice.
   • Click Connect, then Send test event: a real event lands in your Sentry Issues feed.
   • Portal errors then report themselves as they happen: browser crashes and server-route failures, tagged with your shop.
   • The scope is the portal only: your admin is not tracked, and storefront theme errors are filtered out.

   

   Checkpoint: the test event is in your Sentry Issues feed.

   This watches your subscribers’ side, not the machine. Monitoring for the admin, background jobs, and webhooks is a deployment-level setting for whoever runs your code; the platform also watches itself: When an alert opens.

Operational health keeps an eye on all three. Each card gets a row under Integration connections on Settings > Operational health: PostHog and Sentry are re-verified live, Triple Whale shows Connected only, and a failing row names the reason. The panel is part of the morning check.