Mercure
Pub/sub hub pushing updates over Server-Sent Events with Last-Event-ID replay.
- Provider
- default
- Category
- Edge
- Official docs
- https://mercure.rocks/spec
| property | type | values | required | kind default | this component |
|---|---|---|---|---|---|
| ackIntervalSeconds | integer | — | no | — | inherited |
| autoscaling | boolean | — | no | false | inherited |
| channelAssignment | enum | random | hash | sticky | no | random | inherited |
| drainOnScaleIn | boolean | — | no | false | inherited |
| jwtAuthorization | boolean | — | no | true | inherited |
| managed | boolean | — | no | false | ▸ false |
| protocol | enum | websocket | sse | long-poll | grpc-stream | udp-multicast | no | websocket | ▸ sse |
| pushFallback | boolean | — | no | false | inherited |
| sequenceNumbers | boolean | — | no | false | ▸ true |
| stampedeGuard | boolean | — | no | false | inherited |
| stateful | boolean | — | no | true | ▸ true |
| subscriptionsPerUser | integer | — | no | — | inherited |
| topicMatcher | enum | exact | urlpattern | no | exact | inherited |
Every rule whose selector covers this component's kind or its category. Whether a rule fires depends on the design: its properties, its edges and its groups.
When a gateway holding millions of sockets restarts, every client reconnects at once and the gateway asks the coordination service about each of them. Without a stampede guard that lookup storm takes the coordination service down too.
- — all of these hold:
- — the node has at least one synchronous outgoing edge to a coordination-service
- — the node's `stampedeGuard` is false
With heavy-tail traffic a few channels carry most of the load. Assigning channels at random lands those hot channels on arbitrary hosts, and every tenant sharing a host suffers for it.
- — all of these hold:
- — the node's `channelAssignment` is "random"
- — the design requirement `traffic.skew` is "heavy-tail"
Scaling a gateway down kills every live WebSocket on the removed instance at once. Without a drain step those clients all reconnect together.
- — all of these hold:
- — the node's `autoscaling` is true
- — the node's `drainOnScaleIn` is false
nodes:
- { id: "mercure", component: "mercure" }# yaml-language-server: $schema=../../schema/component.schema.json # Mercure: pub/sub hub pushing updates to clients over Server-Sent Events. The Mercure specification # requires subscribers to reconnect with `Last-Event-ID`, on which the hub replays the updates # published after that identifier, and requires a JWT for private updates. Topic matchers are # `exact` by default, with `urlpattern` as the wildcard alternative. schemaVersion: "1" type: "component" id: "mercure" kind: "realtime-gateway" provider: "default" docs: "https://mercure.rocks/spec" defaults: protocol: "sse" stateful: true sequenceNumbers: true managed: false properties: jwtAuthorization: type: "boolean" default: true topicMatcher: type: "enum" values: ["exact", "urlpattern"] default: "exact"