# Weak Signal Finder **Level:** 1 hour ยท your first real analysis **APIs used:** GET /api/v2/net_devices/, GET /api/v2/net_device_signal_samples/ --- Read https://cradlepoint.ericsson.com/llms.txt and the OpenAPI document at https://cradlepoint.ericsson.com/developer/api/openapi.json so you know the Ericsson NetCloud Manager (NCM) API. Build me a Python script (standard library only) that finds the routers with the worst cellular signal over the past 7 days. It covers every router that has at least one cellular modem. Approach: - List cellular modems with GET /api/v2/net_devices/?mode=wan&is_asset=true&expand=router&limit=500. `mode` only accepts wan or lan (mode=cellular silently returns nothing); is_asset=true is what selects modems. expand=router puts the router (id and name) inline. Skip modems whose router is null; they aren't attached to a router. - Pull samples from GET /api/v2/net_device_signal_samples/ with net_device__in=, created_at__gt=, limit=500 and fields=net_device,created_at,rsrp,sinr. The API reference caps a single query at 100 devices, so batch the modem IDs in groups of 100. - Paging: this endpoint has no offset. Follow meta.next exactly as returned until it is null. If a next URL is missing any of your parameters (net_device__in, created_at__gt, fields, limit), add them back before requesting it. If you set order_by, use only created_at_timeuuid or -created_at_timeuuid; ordering by created_at or by a metric makes the API return an error. - Each sample's net_device is a URL; take the modem ID from its last path segment. Pool the samples of all a router's modems together. - For each router, compute the sample count, plus average and 5th-percentile RSRP and SINR, using linear interpolation for the percentile. Skip null readings in the maths, but report how many there were. Rank by 5th-percentile RSRP, lowest first; routers whose RSRP readings are all null go last. Print the 20 worst as a table. Requirements: - List routers that reported zero samples separately, below the table, so "no data" is never mistaken for "good signal". - v2 paths keep the trailing slash. Auth is four headers (X-CP-API-ID, X-CP-API-KEY, X-ECM-API-ID, X-ECM-API-KEY) read from the environment variables X_CP_API_ID, X_CP_API_KEY, X_ECM_API_ID and X_ECM_API_KEY. Send a descriptive User-Agent header. - meta.next may point at a regional host such as www.us0.cradlepointecm.com. Send the same headers there, but only ever send credentials to https:// URLs on cradlepointecm.com or its subdomains. - Don't follow redirects automatically; treat any 3xx response as an error (a 301 usually means a missing trailing slash). HTTP clients re-send custom headers to wherever a redirect points, which would leak your keys. - This is the starter project most likely to hit the rate limit on a large fleet. On HTTP 429, wait until X-RateLimit-Reset (a Unix epoch) if that header is present, otherwise back off exponentially (1s, 2s, 4s, 8s). Make at most 5 attempts per request in total, and never wait more than 60s at a time. On any other HTTP error, stop with a clear message. Use a 30s request timeout. Then tell me how to run it, including which environment variables to set.