Method and apparatus for automatically resolving headroom and service level compliance discrepancies
A workload planner is used to determine headroom and SLE (Service Level Expectation) compliance of a storage system. The historical workload of the storage system over a set of monitoring intervals is obtained, and optionally a new workload to be moved to the storage system is added to the historical workload. Headroom compliance for components of the storage system, and service level expectation compliance on a per-storage group basis are determined by the workload planner. If a headroom or service level expectation violation is determined for one or more of the monitoring intervals, a set of rules are applied to selectively remove monitoring intervals from the time series. The redacted time series is then used to determine headroom and service level expectation compliance. If no violations are found, based on the redacted time series, the storage system meets headroom and SLE requirements.
1. A non-transitory tangible computer readable storage medium having stored thereon a computer program for automatically resolving SLE (Service Level Expectation) compliance discrepancies, the computer program including a set of instructions which, when executed by a computer, cause the computer to perform a method comprising the steps of:
retrieve a time series historical workload for the storage system, the time series historical workload having buckets of data where each bucket of data represents operation of a particular KPI (Key Performance Indicator) of the storage system over a respective monitoring interval; and
determining SLE (Service Level Expectation) compliance of the storage system based on the time series historical workload;
if the selected storage group fails SLE compliance:
determining one or more buckets where the storage group failed SLE compliance;
applying one or more exclusion rules to remove one or more of the buckets where the storage group failed SLE compliance to create a redacted time series workload; and
determining storage group SLE compliance for the selected storage group based on the redacted time series workload.
2. The non-transitory tangible computer readable storage medium of claim 1 , further comprising:
retrieving a new workload that is to be added to a storage system; and
creating a combined time series workload by adding the new workload to the time series historical workload for each of the respective buckets;
wherein the step of determining SLE compliance of the storage system is based on the combined time series workload.
3. The non-transitory tangible computer readable storage medium of claim 1 , wherein one of the exclusion rules is that if the storage system operation exceeds a response time threshold in a respective bucket each week, the respective bucket is removed from the time series historical workload to create the redacted time series workload.
4. The non-transitory tangible computer readable storage medium of claim 1 , wherein one of the exclusion rules is that if the storage system operation exceeds a response time threshold in up to twelve consecutive respective buckets each week, the up to twelve consecutive respective buckets are removed from the time series historical workload to create the redacted time series workload.
5. The non-transitory tangible computer readable storage medium of claim 1 , wherein one of the exclusion rules is that if the storage system operation exceeds a response time threshold in up to two consecutive respective buckets each day, the up to two consecutive respective buckets each day are removed from the time series historical workload to create the redacted time series workload.
6. The non-transitory tangible computer readable storage medium of claim 1 , wherein one of the exclusion rules is that if the storage system operation exceeds a response time threshold by more than 50% for a set of consecutive buckets, but has not exceeded the response time threshold for at least two days, the time series historical workload is reset to begin at the most recent bucket where the storage system response time did not exceed the response time threshold to create the redacted time series workload.
7. The non-transitory tangible computer readable storage medium of claim 1 , wherein one of the exclusion rules is that if the storage system operation exceeds a response time threshold in up to two unrelated buckets in the preceding two weeks, the up to two unrelated buckets are removed from the time series historical workload to create the redacted time series workload.
8. The non-transitory tangible computer readable storage medium of claim 1 , further comprising outputting the determined SLE compliance for the selected storage group.
9. The non-transitory tangible computer readable storage medium of claim 1 , further comprising iterating the step of determining SLE compliance of the storage system based on the time series historical workload for multiple storage group and multiple key performance indicators.
10. A non-transitory tangible computer readable storage medium having stored thereon a computer program for automatically resolving headroom compliance discrepancies, the computer program including a set of instructions which, when executed by a computer, cause the computer to perform a method comprising the steps of:
retrieve a time series historical workload for the storage system, the time series historical workload having buckets of data where each bucket of data represents operation of a particular KPI (Key Performance Indicator) of the storage system over a respective monitoring interval; and
determining headroom compliance of the storage system based on the time series historical workload;
if the headroom of a given component of the storage system is determined to 0.0 TB:
determining one or more buckets where the headroom of the given component of the storage system was determined to be 0 TB;
applying one or more exclusion rules to remove one or more of the buckets where the headroom of the given component was determined to be 0 TB to create a redacted time series workload; and
determining headroom compliance of the storage system based on the redacted time series workload.
11. The non-transitory tangible computer readable storage medium of claim 10 , further comprising:
retrieving a new workload that is to be added to a storage system; and
creating a combined time series workload by adding the new workload to the time series historical workload for each of the respective buckets;
wherein the step of determining headroom compliance of the storage system is based on the combined time series workload.
12. The non-transitory tangible computer readable storage medium of claim 10 , wherein one of the exclusion rules is that if the storage system operation exceeds a best practice limit in a respective bucket each week, the respective bucket is removed from the time series historical workload to create the redacted time series workload.
13. The non-transitory tangible computer readable storage medium of claim 10 , wherein one of the exclusion rules is that if the storage system operation exceeds a best practice limit in up to twelve consecutive respective buckets each week, the up to twelve consecutive respective buckets are removed from the time series historical workload to create the redacted time series workload.
14. The non-transitory tangible computer readable storage medium of claim 10 , wherein one of the exclusion rules is that if the storage system operation exceeds a best practice limit in up to two consecutive respective buckets each day, the up to two consecutive respective buckets each day are removed from the time series historical workload to create the redacted time series workload.
15. The non-transitory tangible computer readable storage medium of claim 10 , wherein one of the exclusion rules is that if the storage system operation exceeds a best practice limit by more than 50% for a set of consecutive buckets, but has not exceeded the best practice limit for at least two days, the time series historical workload is reset to begin at the most recent bucket where the storage system response time did not exceed the best practice limit to create the redacted time series workload.
16. The non-transitory tangible computer readable storage medium of claim 10 , wherein one of the exclusion rules is that if the storage system operation exceeds a best practice limit in up to two unrelated buckets in the preceding two weeks, the up to two unrelated buckets are removed from the time series historical workload to create the redacted time series workload.
17. The non-transitory tangible computer readable storage medium of claim 10 , further comprising outputting the determined headroom compliance for the selected storage group.
18. The non-transitory tangible computer readable storage medium of claim 10 , further comprising iterating the step of determining headroom compliance of the storage system based on the time series historical workload for multiple storage system components and multiple key performance indicators.