Events page is empty: source Google Calendars emptied & plugin-events ICS fetch failing

[rootsongjc]

Summary

All event-related content has disappeared from the site: the Events page is empty, the source Google Calendars are empty, and the Discord events are gone too.

How the Events page is fed

The Events page is populated entirely at build time by the plugin-events plugin (src/plugins/events/index.js), which fetches two Google Calendar ICS feeds configured in docusaurus.config.js:

• Community Call — …/ical/4eef0c86…/public/basic.ics
• Conferences — …/ical/b04a8d52…/public/basic.ics

On a fetch failure the plugin just console.warns and skips that source (returns nothing), so the page ends up empty.

Evidence — two separate problems

1. The source Google Calendars have been emptied. Fetching the ICS feeds directly (curl, HTTP 200):

Calendar	Size	VEVENTs	Notes
Conferences	200 bytes	0	Empty VCALENDAR, only X-WR-CALNAME:HAMi Conference Calendar
Community Call	1113 bytes	1	Only HAMi with CNCF batch-projects (a biweekly sync)

So even when the fetch succeeds, there is essentially nothing left to render.

2. The plugin's ICS fetch is also failing at build time. Build logs show:

plugin-events: failed to parse "Community Call" (https://calendar.google.com/calendar/ical/4eef0c86…/basic.ics), skipping: fetch failed
plugin-events: failed to parse "Conferences" (https://calendar.google.com/calendar/ical/b04a8d52…/basic.ics), skipping: fetch failed

The same URLs return HTTP 200 via curl, so node-ical's fetch is failing where a normal HTTP client succeeds — worth investigating (encoding of the %40, the node-ical HTTP client, or build-env network).

Either way the net effect is identical: plugin-events returns [], so the Events page renders no upcoming events.

Impact

• Events page (/events): empty.
• Google Calendar (source / embed): empty.
• Discord events: also reported empty (likely maintained from the same source/calendar).

Ask

@fishman — you originally set up these calendars and the events integration. Could you take a look:

1. Were the Conferences / Community Call Google Calendars emptied, or did their "Make this calendar public" / sharing settings change? Please restore the events.
2. Any idea why node-ical fails to fetch in the build even though the ICS URLs are reachable via curl?

[Akashstztt]

Investigated the node-ical fetch failure locally.

Both curl and node-ical succeed against the same URL from my machine:

• curl -sI <url> → HTTP 200 OK
• node-ical's fromURL() → resolves successfully, with 2 entries returned.

So the fetch does not reproduce locally. It isn't a %40 encoding issue or a bug in how node-ical builds the request. This points to something specific to the CI environment, such as the GitHub Actions runner's outbound IP being rate-limited or blocked by Google, or an egress/proxy restriction in that build environment, rather than a code-level bug in plugin-events.

@rootsongjc — do you know if this has happened on other CI jobs that hit Google APIs, or if GitHub Actions runner IPs have been flagged before? That would help confirm the CI-network theory.

Regardless of the root cause, I think it's worth making plugin-events more resilient to transient fetch failures by adding retries with backoff and logging the actual HTTP status rather than just the error message. That way, a temporary network hiccup doesn't silently empty the entire Events page. I'm happy to put together a PR for that piece.

Separately, regarding the empty calendar data itself, that's outside what I can fix in code and still needs @fishman to confirm or restore the source calendars.

[rootsongjc]

The calendar item was unexpectedly deleted.

[fishman]

There isn't really much to investigate. One of the people that are shared as admin on that google calendar,deleted the entries.

[Akashstztt]

Thanks @fishman, that clears up the calendar side — sounds like it's now a matter of someone with admin access restoring the entries, so there's nothing further needed from me there.

I'll go ahead with the plugin-events retry-with-backoff improvement mentioned above, since it's a worthwhile resilience fix regardless of the cause of this specific incident. PR coming shortly.

[Akashstztt]

Hi @rootsongjc I’m interested in working on this issue. If it’s available, could you please assign it to me? Thanks!