Describe the working profile
Record concurrent users, 1C configurations, database size and growth, document posting intensity, heavy reports, integrations and scheduled jobs. Pay special attention to peaks such as month-end closing, because they often define the required capacity.
Balance cores and per-core performance
Some operations scale across processes while others remain sequential. Many slow cores are therefore not always better than a balanced CPU with strong per-core performance. Treat the 1C application tier, database and remote sessions as separate roles, even when a small deployment combines them on one host.
Memory and storage control latency
The database uses RAM for data caching, 1C uses it for worker processes and the operating system uses it for file cache. Insufficient memory creates unpredictable disk access. For storage, random-operation latency and resilience matter more than a large sequential-throughput figure. Separate data, logs, temporary files and backups where practical.
Measure before purchasing
Collect CPU utilisation by core, memory pressure, storage latency, database waits, long queries and locks. For a new environment, run a representative load test against a copy or realistic dataset. The final design must also cover updates, backup, recovery time and growth.
A well-sized 1C server is not the most expensive specification. It is a platform that completes the company’s peak business operation predictably and still has a clear path for expansion.