CloudEventBus circuit breaker stuck in a loop : The circuit is now open and is not allowing calls. #75
Assignees
Labels
app: api
Concerns the API
Comments
|
The error is unhappily due to a payload size restriction on |
|
Not sure, I guess the dead letter queue might be an option. Anything that isn't getting stuck in a loop forever is better that the current behavior I think. |
|
I tend to disagree. My main focus was to ensure reliable and ordered delivery of outbound cloud events. I personnally don't see the harm in being stuck in the loop (it's what it's intended for), even though the queue might quickly become abolustely huge in case of delivery failures. @tsurdilo @antmendoza Do you guys have any idea regarding that issue? |
|
no idea @cdavernas sorry. but the error is weird, that big is the payload? I have not been able to see the error in my environment, it is failing before I think, do I have to provide any input to the workflow? |
|
sorry, I was running the previous version of synapse, now the error is different, pasting here just in case you find it useful. |
|
@antmendoza you might need to create the user first on the petShop swagger demo api (https://petstore.swagger.io/#/user/createUser) and use a payload similar to this one to start an instance: {
"username": "test",
"password": "test",
"quantityToOrder": 1
} |
|
@antmendoza In addition, I see that whenever a worker exits, even successfully, it does not seem to gracefully shutdown the GRPC connection, thus resulting in the first set of errors. The second set, though, is due to a bad routing configuration for the error endpoint. |
Yes, in some case, like when taking the unfiltered result of Swagger's Petstore findPetByStatus action, the payload can (but really shouldn't) be considerable. And, as a free service, The question, here, is: do we want to enforce proper and ordered delivery of I personally prefer first option as the dead-letter queue will anyway grow continuously if/when/while facing delivery failures. |
|
Closed as fixed in #366 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
What happened:
The following error keeps appearing in the server logs:
(because the output payload of a workflow/state/action was too big???)
What you expected to happen:
No error, or, at best, a limited numbers of retries?
How to reproduce it:
Not sure, try:
{ "id": "order-dog", "version": "1.0.0", "specVersion": "0.8", "name": "Order pet", "description": "A workflow used to order a pet on the Swagger pet store", "functions": [ { "name": "get-user-by-name", "type": "rest", "operation": "https://petstore.swagger.io/v2/swagger.json#getUserByName" }, { "name": "login-user", "type": "rest", "operation": "https://petstore.swagger.io/v2/swagger.json#loginUser" }, { "name": "get-pets-by-status", "type": "rest", "operation": "https://petstore.swagger.io/v2/swagger.json#findPetsByStatus" }, { "name": "create-order", "type": "rest", "operation": "https://petstore.swagger.io/v2/swagger.json#findPetsByStatus" } ], "states":[ { "name": "PlaceOrder", "type": "operation", "actions": [ { "name": "Get User Info", "functionRef":{ "refName": "get-user-by-name", "arguments":{ "username": "${ .username }" } }, "actionDataFilter":{ "toStateData": "${ .user }" } }, { "name": "Login", "functionRef":{ "refName": "login-user", "arguments":{ "username": "${ .username }", "password": "${ .password }" } } }, { "name": "Get Dog To Order", "functionRef": { "refName": "get-pets-by-status", "arguments":{ "status": "available" } }, "actionDataFilter":{ "toStateData": "${ .dogToOrder }" } }, { "name": "Place order", "functionRef":{ "refName": "create-order", "arguments":{ "petId": "${ .dogToOrder.id }", "quantity": "${ .quantityToOrder }", "complete": true } }, "actionDataFilter":{ "toStateData": "${ .order }" } } ], "end": true } ]}Environment:
Win 10 x64 - self hosted
The text was updated successfully, but these errors were encountered: