November 19, 2026
DIA Data Integrity App
DIA Data Integrity App gives Splunk and data administrators detailed control panels for one question: is the data in every index and sourcetype correct and usable by CIM-based content such as correlation searches, dashboards and Splunk Enterprise Security?Built by mehmet ali canpolatSplunk Enterprise, Splunk Cloud
Platform Version: 10.6, 10.5
CIM Version: 8.x
Log in to rate this app
DIA Data Integrity App gives Splunk and data administrators detailed control panels for one question: is the data in every index and sourcetype correct and usable by CIM-based content such as correlation searches, dashboards and Splunk Enterprise Security? Data model mapping, verified: for 28 CIM data models the app runs tstats by index and sourcetype and records which sources really land in each data model. The verdict comes from tstats, not from CIM index macros or tags alone. Acceleration coverage and cim_*_indexes macros are checked too. CIM field profile: for every data model and source the app samples events and measures each CIM field - how often it is filled, how often it only holds placeholders such as unknown or -, and whether its values follow the CIM definition (expected values, number type, port, IP, hash and MAC shapes). Recommended fields that a source never fills are listed per source. Field definitions are read from the data models installed on your search head, so custom models and newer CIM versions are covered automatically. Field anomalies: every extracted field of every source is checked for broken names (generic CEF cs1/cn1 fields with their real name suggested from the label, positional field3 names, values used as field names, case-variant duplicates), malformed values (key=value inside a value, leftover quotes, spaces, unparsed JSON/XML, encoding artefacts), mixed number/text types, mostly empty fields and timestamp fields that cannot be parsed or disagree with _time. Parsing and time: unparsed and partially parsed sources, multi-line and future-dated events, indexing latency and likely time zone offsets. Recommendations for unmapped sources suggest the data model, tags, key CIM fields and the add-on to use. No setup is needed: after install, first-run searches build the analysis automatically within minutes and scheduled searches refresh it every 4 hours. Defaults are light (7 days, acceleration summaries only, 200-300 sampled events per source); every window, sample size and limit is a macro, and the Overview shows the p90 run time of each job with tuning recommendations. The app is read-only for your data and never changes props, transforms, data models or macros.
Log in to report this app listing.