jthomperoo/predictive-horizontal-pod-autoscaler

Errors in phpa operator pod

Closed this issue · 3 comments

MarahAbu commented

Describe the bug
I have these two errors in the phpa operator pod:

  • ERROR failed to get predicted replica count,......,"error": "exit status 1: Traceback (most recent call last):\n File "/app/algorithms/holt_winters/holt_winters.py", line 92, in \n model = sm.ExponentialSmoothing(algorithm_input.series,\n File "/usr/local/lib/python3.8/site-packages/pandas/util/_decorators.py", line 207, in wrapper\n return func(*args, **kwargs)\n File "/usr/local/lib/python3.8/site-packages/statsmodels/tsa/holtwinters/model.py", line 292, in init\n self._initialize()\n File "/usr/local/lib/python3.8/site-packages/statsmodels/tsa/holtwinters/model.py", line 432, in _initialize\n return self._initialize_heuristic()\n File "/usr/local/lib/python3.8/site-packages/statsmodels/tsa/holtwinters/model.py", line 447, in _initialize_heuristic\n lvl, trend, seas = _initialization_heuristic(\n File "/usr/local/lib/python3.8/site-packages/statsmodels/tsa/exponential_smoothing/initialization.py", line 59, in _initialization_heuristic\n raise ValueError('Cannot compute initial seasonals using'\nValueError: Cannot compute initial seasonals using heuristic method with less than two full seasonal cycles in the data.\n"}
  • ERROR failed to update PHPA configmap

Predictive Horizontal Pod Autoscaler Version
The version of the Predictive Horizontal Pod Autoscaler the bug has been found on.
v0.11.1

To Reproduce
Steps to reproduce the behavior:

  1. Deploy latest phpa version (v0.11.1)
  2. Run with holt_winters model
  3. See error

Expected behavior
A clear and concise description of what you expected to happen.
num of pods being predicted after two-three load tests without the errors described above

Kubernetes Details (kubectl version):
Kubernetes version, kubectl version etc.
v1.20

Additional context
Add any other context about the problem here.

amitde69 commented

@jthomperoo any update on this one ?

jthomperoo commented

Hi @amitde69 I've identified the issue as being with the Python logic that runs the holt winters calculation - I just haven't had time to address any of the issues in this project for a while since I've been very busy. I'll try to make some time to have a look when I can.
Thanks!

jthomperoo commented

This fix is now available in v0.11.2, please let me know if your issue is not resolved.

  • 👍1