1) What is the standard best practice for retaining IH Data -1 month, 3 month or 6 months ? (approx customers~ 2million, Inbound web and mobile channel only)
2) In above scenario if OOTB strategies are used in CDH 8.5 with zero customizations done on strategies, does having a large IH volume (~ 150 million records) adversely impact application performance ? If yes what should be the optimal IH volume after which IH records should be archived ?
3) Adaptive models use ADM datamart tables and does not depend on IH data, hence cleaning IH data does not impact ADM models. Is this understanding correct ?
I saw a company with 20M+ customers and 13+ channels, and they were using 3 months. But maintenance was really hard and I cannot say everything was smooth. But I think in your case 3 months should work smoothly. But your business cases only require 45 days at most, I think you should go with 45 days, rather than Pega's max. capability. Since import is time-consuming, each day counts.
For ADM, cleaning IH tables is not going to harm the adaptive models, but not sure about the pxDecisionResults.