Hosting Service Levels and Backups: Questions to Ask Before Buying

Start with the consequences of interruption

Describe one realistic interruption for your project: visitors cannot read a catalogue, staff cannot process enquiries, or recent submissions disappear. Identify which functions must return first. Google's disaster-recovery guidance separates recovery time objective, the acceptable outage duration, from recovery point objective, the acceptable period of data loss. Use those concepts to express your needs without inventing provider guarantees.

Ask each candidate for the agreement covering the exact plan. Record its availability commitment, recovery commitments if any, and the document date. A target in a sales conversation may not be a contractual remedy. Leave unsupported numbers blank and identify the consequence for your project if the provider cannot confirm a requirement you consider essential.

Read how availability and remedies are defined

Which components count: network connectivity, the server, the hosting platform, or your complete application? Ask where monitoring runs, which reporting period is used, and what qualifies as downtime. Read exclusions for scheduled maintenance, customer changes, and upstream events. Compare the definition before comparing percentages; two apparently similar targets may cover different parts of the visitor experience.

Ask how to submit a service-credit claim, its deadline, required evidence, and any credit cap. Clarify whether the remedy is account credit or another form of compensation. Support response, problem resolution, and backup restoration are separate commitments. This guide establishes no Hightide uptime history or SLA result; confirm applicable terms before treating any target as a promise.

Turn the backup label into a coverage record

Request a list of protected items: website files, databases, uploaded media, mailboxes, and configuration. Ask how often copies are made, how many versions are retained, and whether database and file copies are consistent with each other. Clarify exclusions for large files or storage limits. A snapshot description alone does not answer whether recent application records can be recovered.

Ask who monitors backup failures and who can authorize restoration. Confirm whether restores replace the live service or can be inspected separately, and whether retrieval or restoration incurs a fee. Cloudflare documents configurable caching behavior; cached public assets do not establish a complete application backup. Keep cache availability and recoverable customer data as separate questions in the comparison.

Confirm location, access, and retention boundaries

Request the locations of the running service and backup copies separately, including any selectable region and relevant subcontractor information. A country-code domain does not establish either location. Ask what evidence supports a required residency condition, who can access copies, and how retention changes after cancellation, nonpayment, or suspension. Missing location evidence should remain an unresolved procurement condition.

Agree who maintains an independent export and what formats make it usable elsewhere. Ask whether account suspension prevents downloading existing backups and how long access remains after termination. Review encryption and access controls through the provider's published explanations without inferring zero logging. Keep the backup inventory available outside the hosted application so an outage cannot hide your recovery notes.

Request a practical support and restore plan

Hightide paid activation requires manual review; arrange migration around confirmed readiness. Use dashboard tickets to request support hours, escalation, restoration scope, timing, and charges for the selected plan. If a free isolated restore demonstration is available, use non-sensitive sample data and record the actual outcome. Otherwise mark restoration capability unverified and resolve that gap before entrusting critical data.