I have a dashboard that a specific team uses. Today, they asked about why one of the panels was broken. Looking into it, we were receiving this error from the search:
Error in 'fit' command: Error while fitting "StateSpaceForecast" model: timestamps not continuous: at least 33 missing rows, the earliest between "2024-01-20 07:00:00" and "2024-01-20 09:00:00", the latest between "2024-10-02 06:00:00" and "2024-10-02 06:00:01"
That seemed pretty straight forward, I thought we might be missing some timestamp values. This is the query we are running:
|inputlookup gslb_query_last505h.csv | fit StateSpaceForecast "numRequests" holdback=24 forecast_k=48 conf_interval=90 output_metadata=true period=120
Looking into the CSV file itself, I went to look for missing values under the numRequests column. We have values for each hour going back for almost a year. The timestamps mentioned in the error look like:
Looking at that SS now, There is an hour missing there. The timestamp for 08:00.
That may be the cause. How would I go about efficiently finding the 33 missing values? Each value missing would be in-between any two hours. Will I have to go through and find skipped hours among 8k results in the CSV file?
Thanks for any help.
