// free tool, no account
SIP SRV and NAPTR lookup.
Enter a SIP domain. We walk the RFC 3263 chain the way a phone would: NAPTR, the SRV records for UDP, TCP and TLS (both _sips._tcp and the older _sip._tls), then every target's addresses. Then we tell you what is wrong with it.
a domain, or paste a URI like sip:alice@example.com · queried live via 1.1.1.1
DNS is right today. Will it be right at 3 AM?
Records break when someone renames a host, moves an SBC or lets a zone lapse, and nothing on the server tells you. SIP/SENTRY runs this exact check on a schedule, along with real OPTIONS, REGISTER and TLS checks, and alerts you when a target stops resolving. Three monitors are free.
How SIP finds a server
How does a SIP client find a server from a domain name?
RFC 3263 defines the order. The client first asks for NAPTR records on the domain, which say which transports the domain offers (SIP+D2U for UDP, SIP+D2T for TCP, SIPS+D2T for TLS) and in which order of preference. Each NAPTR points at an SRV name such as _sip._udp.example.com. The SRV records list the actual servers with a priority, a weight and a port. Finally the client resolves each SRV target to A or AAAA addresses. If there are no NAPTR records the client queries the SRV names directly, and if there are no SRV records either it falls back to the domain's own address on port 5060, or 5061 for SIPS.
Which SRV records does a SIP domain need?
At minimum one SRV record for every transport your devices use: _sip._udp for UDP, _sip._tcp for TCP and _sips._tcp for SIP over TLS. Many carriers and phones still use the older name _sip._tls for TLS, so publishing both costs nothing and reaches both kinds of client. Each record names a target host and port. More than one target per transport, at the same priority for load sharing or a lower priority for standby, is what gives you failover: a client that loses its first target moves to the next one without anyone changing a configuration.
What does a broken SIP SRV record look like?
The usual faults are an SRV target that no longer resolves because the host was renamed or removed, a NAPTR record that advertises a transport whose SRV record is missing, a single target with no failover, and an SRV record whose target is a CNAME, which RFC 2782 forbids and some clients refuse. None of these show up in a ping or a port check, because the server itself is fine: only the route to it is broken. Phones that resolve DNS on every registration fail over to nothing, or fail outright.
Why would this lookup show something different from my phone?
This tool asks a public resolver (1.1.1.1) at the moment you press the button, so it sees what the internet sees. A phone or PBX may still hold an older answer in its cache until the TTL runs out, or use an internal resolver that serves a split-horizon view. If the records here are right and a device still misbehaves, compare the TTL with the time since you made the change.