Breakpad vs NHN Cloud Log & Crash Search

Breakpad is a free library and tool suite for capturing application crashes in compact minidump files and converting them into readable C and C++ stack traces. Its client can write a dump when a crash happens or when an application requests one. A symbol dumper converts compiler debugging information, including DWARF, STABS, and Windows PDB files, into Breakpad symbol files; a processor combines those symbols with dumps to produce traces. The sender can transmit dumps and application-defined key-value parameters to a crash server in a multipart HTTP POST request, with HTTPS supported. Developers provide the server and keep symbol files where the processor can access them. Documentation describes integration on Windows, macOS, and Linux, with Android also listed among the platforms. Windows out-of-process dump generation needs a separately run crash generation server. On Linux, crash callbacks should do little because application state may be unsafe; the client design also cautions against heap use and resource allocation after an exception. The project is hosted in the Chromium source repository and lists separate discussion and development mailing lists. The free plan costs nothing.

NHN Cloud Log & Crash Search collects and stores logs from client and server applications for viewing and searching. Logs can be sent to its collector over HTTP as JSON, including bulk delivery in JSON-array format. The Log Search API accepts Lucene queries and requires a project secret key in the request header. Searches cover logs from the previous 90 days, and each time range can span no more than 31 days. Results are returned in pages of up to 100 logs; the guide documents paging through up to 100,000 logs. HTTP ingestion requests are limited to 52 MB, while an individual JSON log can be 8 MB. Search usage consumes tokens based on duration, size, and query complexity, with replenishment over time. Source maps are supported. Platforms listed are Android, iOS, web, Windows, and API. Pricing is by custom estimate rather than a stated product price.