WhatsApp bots intermittently experience delayed webhook delivery. Some incoming messages trigger the webhook immediately, but others are delayed by 5 minutes to several hours before arriving. This makes real-time conversational bots unusable — users tap a button or send a photo, get no response, then receive a random reply hours later.
The issue disproportionately affects:
Image messages
WhatsApp Flows responses
List message replies
Button message replies
Plain text messages seem less affected.
What I’ve verified (this is NOT an n8n or infrastructure issue)
Traefik access logs: Every webhook POST from Meta that arrives receives a 200 OK in 9-43ms. Zero 429s, 502s, or 503s. Traefik is not rejecting or slow-responding to anything.
n8n debug logs: Every webhook that arrives is processed successfully with zero errors (N8N_LOG_LEVEL=debug, monitoring “Received webhook” entries).
Controlled test: Sent 6 photos in sequence. Only 4 webhooks arrived immediately. The other 2 arrived later via what appears to be Meta’s retry mechanism.
Two independent servers: Same issue on two completely separate VPS instances (different hosting providers, different regions, different n8n versions). This rules out any server-side problem.
WABA subscriptions: Verified via Graph API GET /{WABA_ID}/subscribed_apps — all apps correctly subscribed.
Webhook fields: messages field subscribed, all on consistent API version v24.0.
Rate limits: App rate limit at 3% usage — nowhere near throttled.
Rate limiter removed: Had a Traefik rate limiter configured, removed it entirely — issue persists.
Environment
n8n version: 2.7.5
Database: PostgreSQL 16
Reverse proxy: Traefik v2.10
Running on: Docker (self-hosted)
OS: Ubuntu
Started
Approximately February 10, 2026.
Conclusion
Meta’s webhook delivery pipeline appears to be intermittently queuing some webhook dispatches instead of sending them immediately, then delivering them later via retry. My server is proven clean — every request that arrives gets a 200 OK in under 50ms. I’ve filed a bug report with Meta Developer Support. Is anyone else experiencing delayed WhatsApp webhook delivery recently (Feb 2026)? Especially with interactive message types like Flows, Lists, Buttons, and Images? I’d like to know if this is a wider platform issue.
It seems to be a general issue. I’ve spent two days running tests with the tools, the server, etc., and in the end, WhatsApp is the only thing left. I’ll open a ticket tomorrow to see if there are any specific details.
hey @reuben, yeah this is 100% a Meta-side issue and you’ve done all the right diagnostics to prove it. I’ve been dealing with the exact same thing on a client project since around the same time, early February, and it’s maddening because there’s literally nothing you can do on your end to fix it.
The pattern you’re describing with images and interactive messages being hit harder than plain text matches what I’m seeing too, my theory is that those message types go through additional processing on Meta’s side (virus scanning for media, flow validation, etc) and whatever queue or service handles that is either overloaded or has a bug in the retry logic. The fact that they eventually arrive via retry tells you Meta’s system is treating them as failed deliveries even though your server never saw the first attempt.
@Thoth_AI@julio_dominguez@gibacache@aseazul would be worth everyone chiming in on the same Meta developer support thread if you can find it, or opening your own tickets and referencing each other’s case numbers. Meta tends to ignore individual reports but when multiple developers report the same issue with similar timing they actually look into it. If anyone gets a case number from Meta support share it here so others can reference it in their tickets.
Hostingers been having issues too, I wonder if they are both connected.
EDIT - I checked the hostinger status and its having a status page issue, im only posting this here too since a lot of people are encountering hostinger issues.
I have heard back from our admin team and it seems that it’s related to an ongoing issue with Meta. We are aware of this issue and can confirm it is caused by Meta rate-limiting traffic to our network (ASN 47583) on their side. This blocking occurs within Meta’s fwdproxy before requests reach your VPS, meaning we cannot resolve it via server-side configuration.
Our engineers are actively contacting Meta to request that these limits be adjusted. To expedite a resolution, we strongly recommend that you also open a support ticket with Meta/Facebook. When reporting, please specify that:
Your service is hosted on ASN 47583.
Requests/Webhooks are being rate-limited by Meta’s fwdproxy.
The more reports Meta receives from independent VPS owners, the faster they typically prioritize a fix. You can contact their support here: Facebook Help or Meta for Developers .
We are monitoring the situation closely and will continue pushing for a resolution. Thank you for your patience!
As addressed above, Meta is definitely throttling the API gateway into Hostinger.
The fix for me was move my Meta gateway elsewhere. I set one up on GCP and routed all traffic through there. I also moved the webhook endpoint to a standard n8n webhook that calls the gateway, this stops Meta from defaulting back to the Hostinger WhatsApp webhook, which isn’t working.
This setup has completely solved the problem for me.