6b2c7245de
Armada uses a Kubernetes watch to implement its chart wait logic. This can be a fairly long-lived connection to the Kubernetes API server, and is vulnerable to disruption (if, for example, the kubernetes apiserver chart is being upgraded). This change allows Armada to retry the wait for some specific errors, including the establishment of a new watch, until the overall chart timeout is reached. https://github.com/kubernetes-client/python/issues/972 urllib3.exceptions.ProtocolError: ('Connection broken: IncompleteRead(0 bytes read)', IncompleteRead(0 bytes read)) Change-Id: I3e68a54becadd5b2a2343960a120bdc3de8e8515
36 lines
725 B
Plaintext
36 lines
725 B
Plaintext
deepdiff==3.3.0
|
|
gitpython
|
|
grpcio==1.16.0
|
|
jsonschema>=2.6.0
|
|
keystoneauth1==2.21.0
|
|
keystonemiddleware==4.9.1
|
|
kubernetes>=11.0.0
|
|
Paste>=2.0.3
|
|
PasteDeploy>=1.5.2
|
|
protobuf>=3.4.0
|
|
PyYAML==3.12
|
|
requests
|
|
retry
|
|
prometheus_client==0.7.0
|
|
|
|
# API
|
|
falcon
|
|
uwsgi>=2.0.15
|
|
|
|
# CLI
|
|
click>=6.7
|
|
|
|
# Oslo
|
|
oslo.cache>=1.5.0 # Apache-2.0
|
|
oslo.concurrency>=3.8.0 # Apache-2.0
|
|
oslo.config!=4.3.0,!=4.4.0,>=4.0.0 # Apache-2.0
|
|
oslo.context>=2.14.0 # Apache-2.0
|
|
oslo.db>=4.24.0 # Apache-2.0
|
|
oslo.i18n!=3.15.2,>=2.1.0 # Apache-2.0
|
|
oslo.log>=3.22.0 # Apache-2.0
|
|
oslo.messaging!=5.25.0,>=5.24.2 # Apache-2.0
|
|
oslo.middleware>=3.27.0 # Apache-2.0
|
|
oslo.policy>=1.23.0 # Apache-2.0
|
|
oslo.serialization!=2.19.1,>=1.10.0 # Apache-2.0
|
|
oslo.utils>=3.20.0 # Apache-2.0
|