Буферизация и пакетная отправка

Real devices lose their connection: the router reboots, the signal drops, the cloud blips. A good logger keeps working when the network disappears. 📶

If you only send readings live, an outage means lost samples forever. The fix is to store offline first, then upload once the network returns.

A buffer is a temporary holding area for data waiting to be sent. Your readings queue up there safely until an upload can happen.

For a few values, an array in RAM works. For many that must survive resets, buffer to an SD card file or flash instead.

Batching means sending many readings in one request instead of one at a time. Fewer connections save power and play nicely with the cloud.

The pattern is simple: always save a reading to the buffer, and only try to upload when WiFi is connected.

Before every upload, confirm you are online. Skipping this check makes your code hang or throw errors during an outage.

When online, package the queued readings, often as a JSON array, and POST them together in a single batch request.

Delete buffered data only after the server confirms it. If the upload fails, keep the data and retry next time.

A growing buffer can overflow RAM. Cap its size, or write to storage, so a long outage never crashes the device.

If uploads keep failing, wait longer between tries. This backoff saves battery and avoids hammering a struggling server. 🔋

When should you delete readings from your offline buffer?

Always save first, upload in batches when online, and clear the buffer only on confirmed success. Your logger now survives any outage. ✅