Struggling to get a successful deployement on the first deployement

Hi team!

it’s getting more difficult to get a first deployement successful. I often have to “restart” or “re deploy”. But over the last 2 weeks, it’s getting really difficult, example, for the “same” commit (sometime just changing a word just to trigger a deployement):

Right now I cannot deploy since yesterday (while building successfuly locally) I don’t know what do do - what do you recommend ?

Hello @CBaptiste,

If you look at your deployment logs, you can see that the liveness & readiness probe checks are failing:

And in your application logs, there seems to be an issue at runtime:

I can see this error from at least yesterday around 23:36 until now

Hey @CBaptiste !

To complement @Melvin_Zottola answer, it seems your app cannot be scheduled for every new deployment because of lack of resources on your cluster and a new node has to be provisioned for each deploy, CF deployment logs:

Looking at your settings, your cluster is set from 3 up to 10 nodes (you actually have 5 running) nodes have 2 CPU and 4 Go RAM, your app is requiring 1500 milli CPU and 2Go RAM, so one node can fir only on instance of your app. Meaning every time a new instance of your app is needed (deployment or scaling) then a new node has to be provisioned, this process can take several minutes depending on instance type / region.

Is there any reason this app need so much CPU? You will probably benefit lowering its CPU if not required to 500milli so at least 2 instances can fit on the same node AND / OR take bigger AWS instances.

Changing a bit the settings and / or increasing instances size will probably speed up deployment.


Thank you! Will try a bigger AWS machine :slight_smile: