关于Firebase控制台推送通知TPS限制及调优的技术咨询

Hey there, let's tackle your two questions clearly based on Firebase's capabilities and real-world usage:

1. What's the TPS limit for Firebase Push Notifications? #

The transaction per second (TPS) limit for Firebase Cloud Messaging (FCM) depends on your pricing plan:

• Free Spark Plan: The default TPS cap is 500 requests per second. This applies to both API calls from your server and notifications sent via the Firebase Console, since the console uses the same underlying FCM APIs.
• Blaze Pay-as-you-go Plan: The default TPS limit jumps to 10,000 requests per second. If you need even higher throughput, you can submit a request to Firebase support to increase this limit—they'll review your use case and adjust it accordingly (usually for high-volume legitimate use cases).

It's worth noting that these limits are for the FCM service itself, not just the console. So whether you send notifications via the web UI or your own backend, you'll hit these caps if you exceed the rate.

2. Can I adjust the TPS limit for the Firebase Console? #

You can't directly tweak a "TPS slider" in the Firebase Console to set a custom rate, but there are workarounds to control how fast notifications are sent:

• Batch your sends: Instead of targeting all 1 million users in one go, split your audience into smaller segments (e.g., 10 batches of 100k users) and send notifications to each batch with a time gap (like 5-10 minutes apart). This effectively reduces the effective TPS hitting your server as users come online gradually.
• Upgrade to Blaze Plan: As mentioned earlier, the Blaze Plan has a higher default TPS limit, and you can request further increases. This gives the console more headroom to send notifications faster if you need it, but you still have control over pacing via segmentation.
• Switch to server-side sending: If you need precise control over TPS, use FCM's server APIs (like HTTP v1 or Legacy API) instead of the console. You can implement rate-limiting logic in your backend code to cap the number of requests sent per second, aligning it with your server's capacity to handle the incoming user traffic.

Extra Tip for Your Server Issue #

Since your server struggled with the sudden surge of active users after sending notifications, remember: FCM's TPS limit controls how fast notifications are delivered, but you also need to prepare your own infrastructure for the traffic spike. Consider implementing strategies like auto-scaling server instances, adding caching layers, or using queue systems to handle user requests asynchronously.

内容的提问来源于stack exchange，提问作者ankit.vishen