Hello. I've got a question regarding can-i-deploy step. I'm using it 2 times in my CI/CD workflows. One - when I verify whether I can merge the code, in this case I verify pact against provider's master branch, the other one - when I verify whether I can deploy the code, in this case I verify pact against provider's env. So my script looks this way:

#!/usr/bin/env bash

echo "running can-I-deploy with the following params""
PACT_BROKER_BASE_URL: $PACT_BROKER_BASE_URL
PACT_BROKER_TOKEN: $PACT_BROKER_TOKEN
consumer: $consumer
version: $version
provider: $provider
to_environment: $to_environment
"""

if [ -z "$version" ]; then
  echo "ERROR: version cannot be empty to run can-i-deploy verification"
  exit 1
fi

command=
if [ -z "$to_environment" ]; then
  command="--pacticipant $provider --branch master"
else
  command="--provider $provider --to-environment $to_environment"
fi
echo "command = $command "
docker run --rm \
  -w "${PWD}" \
  -v "${PWD}":"${PWD}" \
  -e PACT_BROKER_BASE_URL="$PACT_BROKER_BASE_URL" \
  -e PACT_BROKER_TOKEN="$PACT_BROKER_TOKEN" \
  pactfoundation/pact-cli:0.50.0.32 \
  broker can-i-deploy \
  --pacticipant "$consumer" \
  --version "$version" \
  $command

I've found that if I want to use

--branch master

in my command I need to use

--pacticipant

, if I want to use

--to-environment

I need to use

--provider

otherwise it returns an error...Am I missing something here? If not...why cannot it be the same (provider or pacticipant) in both cases?

hmmm something else must be the matter. All of our examples use the format

pact-broker can-i-deploy \
	  --pacticipant ${PACTICIPANT} \
	  --version ${GIT_COMMIT} \
	  --to-environment ${ENVIRONMENT}

Right, the first pacticipant is consumer. What if I want to specify the provider name in the command?

Why are you specifying two

--pacticipant

values in this case? I think that’s your issue

Ok my case is here: we have a multi-tenant consumer with single-tenant providers. So when I run can-i-deploy step, I need to verify pact between the consumer and different providers. Some providers are deployed to all envs, some providers are deployed only to some envs. What I want to make sure when I deploy my consumer is that a 'valid' version of the provider is deployed to env1, env2, env3. I am going to use strategy.matrix for this:

strategy:
  matrix:
    provider: [<provider1>]
    env: [
      "<env1>", "<env2>", "<env3>"
    ]

and I expect my command call to looks this way (in a loop for each env):

broker can-i-deploy \
--pacticipant "$consumer" \
--version "$version" \
--provider "<provider1>"
--to-environment "<env1>"

I'm still not getting why you need to specify the providers in the command tho. by asking can I deploy and giving a specific application and environment, it will tell you if there are missing dependencies, or if those dependencies are incompatible.

ok lets I have a consumer C and it has 2 providers P1 and P2. • C is a multi-tenant service and gets deployed to a single env

prod

. • P1 is a single-tenant service and gets deployed to client1_prod, client2_prod, client3_prod envs. • P2 is a single-tenant service and gets deployed to client1_prod, client2_prod envs I will record deployment for •

P1

envs

client1_prod

,

client2_prod

,

client3_prod

•

P2

envs

lient1_prod

,

client2_prod

Now I need to verify if I can deploy C. I need to verify pacts C - P1 env

client1_prod

C - P1 env

client2_prod

C - P1 env

client3_prod

C - P2 env

client1_prod

C - P2 env

client2_prod

If I don't specify the provider, I need to verify <all pacts for consumer C> for envs

client1_prod

,

client2_prod

,

client3_prod

But verification C - P2 env

client3_prod

will fail, because this service will never be there